Industry Smart Content. THE REAL VALUE OF DATA GOES BEYOND ANALYSIS WITH MICHAEL FORD, AEGIS SOFTWARE SEP/OCT 2019 i40today.com

Connect your company with the Industry 4.0 market that is projected to reach $152 billion by 2020. For more information, visit www.i40today.com. Welcome to Issue 6 of i4.0 Today i4.0 Today is a platform dedicated to the concept, technology & future of Industry 4.0. We are not just another publication we are an interface for sharing all Industry 4.0 news, research and development within electronics manufacturing. Our mission is to Lead the industry 4.0 movement, and give an insight into the concept, technology and future of Industry 4.0 Leading the Industry 4.0 Movement In today’s fast paced and extremely competitive market, every second saved in the workplace saves not only time, but revenue. Using Industry 4.0 technologies to increase productivity and reduce risk is the number one priority emerging with every industry and sector. What to expect from us... • Opportunity to hear industry • Established quarterly publication leading speakers • Industry 4.0 technology events • Dedicated Industry 4.0 website around the world www.i40today.com • Sponsorship opportunities • Global digital campaign Magazine Contacts Editorial & Advertising Enquiries EDITOR DIGITAL MARKETING To enquire about available advertising Wendy Tindle Cameron Shaw opportunities, please contact: [email protected] [email protected] [email protected] GLOBAL SALES WEB & MARKETING Andrew Carpenter Gordon Brown www.i40today.com [email protected] [email protected] NEXT ISSUE 7 CONTENT DEADLINE EVENTS & SALES DESIGN & MARKETING Q4 2nd October 2019 Stacey Bonnar Lorna Hull [email protected] [email protected] 2 An insight into Industry 4.0 Sep/Oct 2019

Inside this issue... 26 Q&A with Drew Elhers, Global Director, Portfolio Marketing 4 The real value of data goes Enterprise Software, Zebra beyond analysis Technologies Aegis Software Zebra Technologies 8 ROBOZE 3D printing technology 30 Making machine-learning and the manufacturing future inference meet real-world Roboze performance demands Xilinx 10 Siemens’ new Digital Enterprise Experience Center (DEX) 34 8 signs your material supports customers worldwide management needs an upgrade Siemens Swissmic 12 The Connectivity Conundrum 36 Minimising the risk of cyber i4.0 Today Guest Writer breaches in the manufacturing industry 16 The factory of the future begins World Wide Technology with the smart factory of today Koh Young 38 IoT and pollution: A breath of fresh air 22 Yamaha delivers configurable Comms365 modular transfer for robot assembly cells 40 Manufacturers are facing an Yamaha unenviable sea of challenges Outsystems 24 “Industrial AI Company” PerfectPattern expands AI 42 Consumers lack trust in artificial competence intelligence PerfectPattern Pegasystems Inc. 44 OMRON contributes to resolving the world’s social issues through its businesses Omron PUBLISHED BY ALL RIGHT RESERVED GET IN TOUCH The content of i4.0 Today is protected by copyright law, full details of which i40 Today Ltd, 1 Dow Road, are available from the publisher. While Prestwick International Aerospace Park, great care has been taken in the receipt Prestwick, Ayrshire, UK, KA9 2TU and handling of material, production Tel: +44(0)1292 834 009 and accuracy of the content in this Email: [email protected] magazine, the publisher will not accept any responsibility for any errors, loss or omissions which may occur. i40today.com 3

The real value of data goes beyond analysis By Michael Ford, Aegis Software Meet Michael Ford, Industry is focusing these days, on the acquisition of data, with fingers crossed i4.0 Today’s new that analytics will in turn, take the data to create value. Industry 4.0, however, is regular columnist about actively optimizing the factory that today has to produce small lot-sizes and a larger mix of products. Continuous optimization is all about looking for, and removing, the waste, including elimination of all forms of down-time. Looking beyond simple internal machine stoppages, Industry 4.0 is all about the analysis of periods of time for which there is no data, as the machine or process is not actually working. How are we supposed to analyze “no data”? The reality is that we need to split the scope of the digital factory into three layers, each of which must come to terms with their own revolution. Layer 1: The Shop-Floor with momentum building rapidly. The concepts behind CFX, including true “plug and play”, the The IPC Connected Factory Exchange (CFX) “last interface you will ever need” and “machine has opened the eyes of people in the digital connections for free”, have really galvanized all manufacturing world. With the massive expense, corners of the industry towards full adoption over resource-drain, risk and deployment-related line the coming months. down-time out of the equation, the “gold-rush” has begun, with more information than ever Layer 2: Smart Manufacturing Software before available from machines, transactional (MES etc.) operations, and people around the shop-floor. The CFX-driven Layer 1 revolution, though published Software systems in manufacturing, for quite recently, is well under way in many sites generations, have been stunted in what they are already, with almost every key machine vendor able to offer and achieve in terms of machine now in the process of publishing roadmaps and connectivity. Solution vendors are now clearly announcing CFX support completion. Many differentiated. Those who played a key role in “Proof of Concept” (POC) projects have started, the creation and development of CFX, who are 4 An insight into Industry 4.0 Sep/Oct 2019

already providing CFX support as standard, The Analog Factory: have current mature solutions based on IIoT architecture, designed to deliver next-generation The production operator, sees a red light on digital manufacturing benefits. Others are playing the machine’s tower, goes to the machine and catch-up, and will put together their CFX support requests the reason for the stop. The machine in the same way as any other interface, a great report shows that it stopped because no product step forward, but only to the extent that the had arrived for processing. The report illustrates original design of their solutions, having been the symptom, but is just a suggestion of what the architected decades ago, allows. No escaping the real problem might be. This is the case in around fact though that every software solution whether 80% of cases with any machine-based report, as internally developed or commercial, has a step- the content often relates to issues originating uplift opportunity with the arrival of CFX. outside of the machine scope. Walking to the preceding machine in the line, from which the The CFX-driven Layer product should have been sent, the operator 1 revolution, though takes out another report. The machine, which published quite recently, is also stopped, shows that it has experienced is well under way in many a material “pickup error”. The operator goes sites already, with almost to the back of the machine and sees that the every key machine vendor feeder causing the problem has been exhausted now in the process of of materials, but there is no replacement reel publishing roadmaps present. OK, simple. Material logistics has and announcing CFX messed up. After hunting around for a logistics support completion. support person, and explaining the situation, the response is that the issue will be looked in Looking beyond the simple interfaces, and to. The kit of materials was prepared in advance focusing more on the IIoT-based MES platform, the day before, everything should be there. After there is critical value being added to the CFX searching the kit, and emptying out the used reels data, that of live contextualization. Let’s consider from the trash, it was determined that one reel of an everyday sequence of events, comparing an materials had in fact been missing from the kit. “analog factory” process against one with digital Checking the kit preparation records, all reels had MES. The process starts as a simple event, an been allocated and included, so, it was assumed SMT machine stops during production. The goal that the missing material must have been stolen is to discover the stoppage, find the cause, restart from the kit – well, strictly speaking, not stolen, production, and optimize the factory operation to but mis-appropriated for use on another line. It make sure that it does not happen again. can happen overnight that another line runs short of materials, due for example, to excess spoilage or an incorrect reel count. Since the warehouse is not working during the night shift, materials are taken from other kits, since that is the only quick option to keep the line running. Rather than fixing the issue, the unexpected shortage of materials is simply propagated from one instance to another. With two machines on stop, the operator is faced with the same pressure to get the line running again. It would take a significant amount of time to issue another reel from the warehouse, accounting everything somehow with ERP, so it was suggested that it may be quicker to go and i40today.com 5

look in the part-used material area, to see whether quality problems occurring weeks later, no there were materials there that would get the line unhappy customer, no loss of business. going in the meantime. After searching, yes, a reel with enough parts was there, with the correct Was this use-case benefit included in the ROI part-number. On to the machine it goes, and calculation for the digitalization of the factory? It production starts. Weeks later in a management is clear that the introduction of digital MES cannot meeting, the number 1 issue is that a customer directly be compared with the old analog factory, has found that unauthorized materials were found as things are happening in different ways. Many in their products, from a manufacturer who had down-stream issues are resolved by root-cause not been qualified. This was discovered as a avoidance of bad practices. The value created result of a defect found in a product at a customer. in this example by the CFX data was created The unauthorized component had a slightly not by the analysis of the data itself, but by the different geometry than the approved parts from operational improvements that it supports. This other vendors, resulting in a wider variation of is just one very simple example of so many positional accuracy, which weakened joints, and common everyday cases that happen in analog caused defects. The production manager vaguely manufacturing, that are eliminated in the digital recalls a drop in the yield at test, and a busy time factory revolution. at repair during the period in which the batch containing the defect was made. Nothing was Layer 3: Artificial Intelligence (AI) done at the time, it was a busy period. Not any more however, as due to this issue, the customer Some operations are familiar with the concepts is now looking for a new EMS partner. of factory process simulation. Not too many though, as the most effective software in this The goal is to discover area is somewhat niche and expensive, and the stoppage, find even that is now obsolete, when set into the the cause, restart Industry 4.0 environment. Simulations are, by their nature, based on timescales of days, production, and optimize weeks or even months of factory operation. The the factory operation to industry 4.0 factory operation changes day by make sure that it does day, in line with immediate customer demand, and so any simulation is worthless almost as not happen again. soon as it is completed. The issue that simulation software addressed however, remains relevant The Digital Factory: to Industry 4.0, and is in fact more critical if the operation is to retain a high productivity, At the start of the above scenario, in the digital whilst also being flexible to support a high mix factory, you would have expected that the of products. The solution however has to be machine stopping in the line would have sent a quite different. Layer 2 MES is quite capable CFX message to the II0T MES system to indicate of managing and actioning schedules, as this that it had stopped due to no arriving product. is part of its fundamental structure, and deals Having knowledge of the line configuration, with what is currently happening on the lines, MES would immediately identify the preceding including all dependencies, making a real-time machine in the line to check on the status. analysis of data in order to automate and assist Unlikely to happen however, as MES would have critical decision-making. Layer 3 on the other already taken action having received the “pickup hand, is focused on finding sense in the areas error” message from the previous machine. In for which there is no data, by understanding fact, this also is unlikely to have happened. MES what is really happening in the operation, in knows the product detail, and how the work is terms of bottlenecks current and near-term, to split between the different machines in the line, optimize out wasted time, or, fill the time with knowing all of the material setup locations. MES added value work, for example the inclusion of is also aware of materials that are loaded to maintenance tasks at times when the line has no the machine, the quantity on each reel, and the active production work to do. A human mind, no number taken per production unit, as well as the matter how experienced or capable, is not able rate at which units were being produced, as well to keep track of the thousands of variables, or as any spoilage. In the lead-up to this scenario, even tens of thousands in some cases, that can MES would have automatically ordered an be used to govern, manipulate and continuously authorized replenishment reel, well ahead of the tune the manufacturing operation. The pinnacle machine stopping. In this case however, wasn’t of Industry 4.0 is the assistance, and eventually a material removed from the kit? MES cannot to autonomy, to optimize high-mix production, predict random human behavior. The digital MES utilizing AI. software however, does not need to use kits, as materials are always allocated and delivered to Conclusion: machines “just in time”. There is never a reason to “steal” materials, there is never a kit shortage, The progression to Layer 3 is enabled by Layers 1 never even a kit. Even if a human operator were & 2, and it is for this reason that layer 3 as yet has to try to utilize a material in an incorrect location, been impossible to even imagine for most people, the system would block it. No chance of resultant and so older increasingly limited simulation technologies endure. Technology for Layer 1 in terms of CFX data capture, coupled with Layer 2 IIoT-based modern MES platform are currently available, and can be put to work immediately to drive incredible shop-floor visibility and control, over the whole of manufacturing. Let’s now look forward to the revolution in solutions at Layer 3. www.aiscorp.com 6 An insight into Industry 4.0 Sep/Oct 2019

Learn more at aiscorp.com Visit us at Booth #628 i40today.com 7

ROBOZE 3D printing technology and the manufacturing future By Ilaria Guicciardini, Marketing Manager, Roboze Additive manufacturing is changing the way products are realized in several industrial sectors. According to Wohlers Report 2019, those companies that have introduced 3D printing systems within their manufacturing processes for the realization of functional prototypes and finished parts represent the 56.3% of the total. The introduction of additive manufacturing within designs and manufactures state-of-the-art the traditional production processes allows a FFF 3D solutions for additive manufacturing company to reduce the lead-time of a product applications. Many industry leaders in Aerospace and better control the quality and the cost of the and Automotive sectors are introducing ROBOZE components in its manufacturing line. Additive 3D printing technologies in their manufacturing manufacturing guarantees also more freedom processes, reducing their production costs and in terms of design and possibility to create time in the most extreme environments. ROBOZE components that could not be easily realized with high temperature super polymers and composite traditional manufacturing because of too long materials are able to meet the most extreme needs production time and extremely high costs. – in terms of mechanical, thermal and chemical properties and Metal Replacement purposes – of That’s here that ROBOZE comes into play. With industries like Aerospace, Motorsport, Oil&Gas the headquarter in Italy and a branch in US, and Automotive. Using polymers instead of the innovative 3D printing company ROBOZE metals is incredibly advantageous in these 8 An insight into Industry 4.0 Sep/Oct 2019

With the headquarter ensures smoothness of the movement, quietness in Italy and a branch in US, and positioning precision equal to 25 micron. The the innovative 3D printing final goal is offering the best industrial 3D printers company ROBOZE designs through a continuous product innovation. The advantages for ROBOZE customers are tangible and manufactures in terms of precision, flexibility, personalization state-of-the-art FFF and accessibility, compared to traditional mechanical processes. 3D solutions for additive manufacturing The industrial sector is therefore witnessing a real revolution, driven by the continuous evolution applications. of additive manufacturing. For companies like ROBOZE, it is extremely important to invest time sectors thanks to the possibility to reduce the and resources in R&D, in order to get aligned weights of the parts and get more workability of with the rapid changes of the market worldwide. the plastics, compared to metal alloys. More and more often, especially in US and Asia, the manufacturing industries use both 3D printing ROBOZE techno-polymers are able to replace and traditional methods with CNC machines. metal alloys, allowing to print functional parts for any application. In the FFF 3D printing market, Carbon In particular, additive manufacturing is changing PEEK is the material with the greatest properties the future of the aerospace and defense sectors: for extreme applications. The addition of carbon here the biggest challenge is reducing the fiber makes the PEEK matrix even more resistant in weights of the aircrafts with lighter parts, keeping terms of mechanics and allows to go further in the high the performance standard of the produced Metal Replacement process in Motorsport, Oil&Gas, components. ROBOZE 3D printing technology – Aeronautics and Aerospace industries. for examples thanks to PEEK, the semi-crystalline thermoplastic polymer that guarantees the most Moreover, ROBOZE managed to reach extreme performance with its mechanical properties extraordinary mechanical precisions in the FFF and the excellent thermal and chemical resistance – 3D printing technology thanks to the innovative allows the companies working in these sectors to Beltless System, removing belts and introducing replace light alloys like aluminium with ROBOZE a direct mechatronic movement of X and Y super polymers, guaranteeing the reduction of axes, with a system of hardened steel racks and weights, fuel consumption and CO2 emissions. pinions. Roboze patented Beltless system also Education and technological know-how, with a good knowledge of printers, materials and post-processing techniques are, therefore, fundamental for this industrial revolution driven by additive manufacturing, that, together with ROBOZE, aims at supporting customers in getting the best advantages from this great opportunity offered by 3D printing. www.roboze.com i40today.com 9

Siemens’ new Digital Enterprise Experience Center (DEX) supports customers worldwide By David Rogers, Global Account Manager, Electronics Industry, Siemens On June 26 and 27 Siemens introduced a “grand new” Digital Enterprise Experience Center (DEX) in Campbell, California. The so-called Silicon Valley DEX will provide tailor-made solutions for electronics manufacturing and semiconductors to support customers in Bay Area and worldwide. With the fast-rising electronics market, Siemens of electronic products itself and learn how to respond recognized the need of a vertical approach which with the digital thread approach. is dedicated to this industry with a global technical center of competence as the center piece in the Promoting Siemens within the electronics market, heart of Silicon Valley. As an expansion of the already Silicon Valley DEX is the only Demonstration exhibited showcases around final assembly & testing Center taking an electronics use case to exercise within electronics, an additional fragment was built the Digital Enterprise story end-to-end for the which starts in the design phase and closes the customers in this area. The shown end-to-end gap to manufacturing, demonstrating an end-to-end scenario starts with the design change of the solution. Through the DEX, customers will be able to Nanobox PC, a Siemens-own product, where collaborate directly with Siemens as a trusted partner necessary simulations need to be made for product facing similar industry challenges as a manufacturer validation. Further, a simulation of the product manufacturing can be observed from factory level 10 An insight into Industry 4.0 Sep/Oct 2019

Through the DEX, customers will be able to collaborate directly with Siemens as a trusted partner facing similar industry challenges as a manufacturer of electronic products itself and learn how to respond with the digital thread approach. to shop floor level, including production layout beginning. The DEX Silicon Valley will furthermore and simulation of the manufacturing process. explore bespoke solutions for our customers via The final setup stage includes a simulation of the value definition workshops. These start from actual Nanobox manufacturing within Camstar, the one day to deep-dive sessions focusing on the preferred Manufacturing Execution System (MES) customers’ dedicated field of interest. And even for Electronics from Siemens. Innovative solutions if customers did not identify the suitable scope of also are shown around the Digital Twin in operation digitalization of their operations yet, we provide via MindSphere and Artificial Intelligence. consulting services to identify sweet spots and formulate a digitalization strategy.” “We believe in massive synergies since the digital thread by Siemens is following the complete Many electronics manufacturing companies value chain in electronics which will provide are already engaging on their path of Industry insights from current operation towards the 4.0. With the Silicon Valley Digital Experience next batch,” said David Rogers, Global Account Center, Siemens established a point of entry for Manager, Electronics Industry. electronics manufacturers despite the step of the value chain. Customers are welcome to visit the He continued, “Covering the full scope of DEX and engage with industry experts on current electronics manufacturing as a first point of industry challenges and solutions. entry is very appealing to customers, but just the www.new.siemens.com i40today.com 11

The Connectivity Conundrum By i4.0 Today Guest Writer One might think there has been way too much talk of connectivity over the last year or two and that the subject is filling way too many column inches. But this is an important issue. In fact, right now it’s among the most important issues related to creating smart factory solutions. Connectivity is the foundation of the digital transformation of manufacturing. Without machine to machine, machine to system and machine to operator communications none of the benefits of industry 4.0 or the smart factory will be realized. To sort hype from fact, we ask four questions to six industry experts who are closely involved in the deployment of smart factory systems. Here’s what they had to say. Discuss the importance of Sam Wong TS - connectivity for Industry 4.0 or IIoT Manufacturing Technology Product James Mok - Strategy, IoT Management Manager at and Big Data at Dassault Keysight Technologies: Systèmes: In the manufacturing world, At the heart of Industry 4.0’s there are many different core enabling technologies vendors and each are lies the mirror image of specialized in their focus area. They have been the physical world, or less concerned about the bigger picture or other sometimes called the Digital equipment, at least not until now. For the industry Twin. These digital twins do not only represent the to optimize their processes, to have higher products but also the processes and operations performance and tap into benefits of Industry4.0, from design, engineering, manufacturing to there must be synergy between these vendors in services. In Dassault Systèmes, we not only terms of data exchange. A common platform or provide solutions for these digital twins, we standards is needed to make this happen. consider it central to our strategy to deliver digital continuity within our 3D Experience platform, Peter Bollinger – CEO of such that these digital twins work seamlessly iTAC Software: with each other. To optimize design, engineering and manufacturing processes, we need to bridge Connectivity is fundamental the gaps between the silos of the value chain, as to implement features well as between the virtual and physical world. and functions of Industry Connectivity is key in delivering digital continuity. 4.0, IIoT or Smart Factory solutions. The goal is that the connectivity is simple 12 An insight into Industry 4.0 Sep/Oct 2019

and standardized but in today’s production What are the main issues or challenges environment we see all kinds of different machine associated with connectivity in our models with different interfaces and controllers industry today? depending on the age. As customers like to improve production quality and overall output of James Mok: the whole plant we need to be able to connect every machine, and that means legacy system There are two main issues. that may have been installed 5 to 15 years ago. Firstly, the standard around At the end Software Solution Providers and communication protocol. machine manufacturer need to work more closer Secondly, the common together to make the connection simpler with semantics to interpret the standardization as customers looking for reliable messages. It would not be a features to drive their goals and are not to much challenge if all the machines concerns about the technical solution to do so. of the world use the same communication protocol and speak the same semantics. There Brain D’Amico - President have been many attempts to establish such a of Mirtec Corp: standard. In the real world of automation, such a dominant standard is almost impossible. This is Connectivity is the critical for several reasons: Firstly, every protocol has its first step toward the own pros and cons, there is always a trade-off in implementation of Industry choosing a protocol depending on the use case at 4.0 within the Electronics hand. Secondly, there are always legacy devices Manufacturing Industry. The and machines that cannot keep up with prevalent ultimate goal is to achieve standards. Thirdly, there is constant development higher levels of productivity by connecting all in the underlying network technology (Ethernet, parts of the manufacturing process through wifi, 5G). Lastly, there is always a political struggle Machine-to-Machine (M2M) Integration and the between powerful vendors and sometimes Industrial Internet of Things. This requires the government bodies to establish standards. collection of data to a centralized platform which Therefore, despite all the efforts to simplify the may then be analyzed. By eliminating inefficient landscape of connectivity standards, it is unlikely procedures, capacity attrition, wastefulness and that one can fit all shapes and sizes. performance bottlenecks, we can optimize the manufacturing process. Sam Wong: Gerry Padnos - Director There are too many vendors, of Technology at Juki each with varying standards. Automation Systems Inc: Some of these are proprietary standards and have different I think of Industry 4.0 as the hardware needs. At other USB of the manufacturing times, customers have world. Everyone wants to legacy systems. For some just plug all their devices Test departments, they may also need to take together and for them directions from their corporate, to align with a global to all “just work” without a lot of hassles or strategy. Corporate will take a longer time to meet frustrations. Having this simple interfacing and the requirements of all their plants, each of which interoperability is going to be very important to has different standards and requirements. achieve continuing efficiency and productivity improvements and reducing costs. It’s pretty Peter Bollinger: clear to us that customers want to have this functionality, but it just hasn’t been available The main issue with so far. We’ve recently seen a lot more pressure standards is that there are on manufacturers to provide simpler and better too many and that legacy connectivity and I think that will continue. equipment will make it very challenging to adopt Thomas Marktscheffel a single standard for any - Director Product factory. In addition to that Management SW- companies have customized business application Integration Platform at that need to be connected as well. Our task is ASM Assembly Systems: to build all the connections to upper and lower levels to execute business processes which vary Smart equipment – like between plants and factories, even when they ASM’s SMT solutions – is belong to the same corporation. What’s more, and remains the basis of we must deal with the existing environment. every efficient electronics factory. Connectivity Companies will not replace their machine or and integration are the new challenges. Only the business applications just so we can implement a electronics manufacturers who provide reliable new standard connection. Each new standard will data in real time and irrespective of time, place be added on top of the other. Over a much longer and device, will be able to run a smart SMT factory time period it may be possible to get to a single with agility and competitiveness. standard, but this would mean that the standard needs to be accepted by all equipment supplier This requires strategic partnerships between in an industrial vertical and survive from technical electronics manufacturers and competent digital perspective for a very long time. transformation experts – for target-oriented operation in teams, for workflows instead of Brain D’Amico: production, and for open standards instead of proprietary islands. Manufacturers and equipment One of the biggest suppliers must “network”– not just technically, but challenges with M2M with customers, partners and competitors. communication is the ability to provide seamless connectivity between such i40today.com 13

a wide variety of assembly equipment. The fact is Peter Bollinger: that different vendors use different communication protocols, data formats, and standards. We continue to maintain a connectivity tool box Gerry Padnos: within our solutions to convert the information we The biggest challenge is get into our own internal getting everyone working standard protocol that toward the same goal. In the feeds all our software modules for seamless and absence of a good standard, reliable communication. Of cause we will add many companies have made new standards as they appear in the market. We their own. Some of these are currently aware of several new standards companies are now reluctant in development, and they are not just those to switch because what they’ve made is working. hitting the headlines, like Hermes or CFX, there There’s also a big challenge with the number of are machine vendors adding new standards or older machines in use around the world. In some updating existing ones too. Cogiscan is part of cases, upgrading these to use new technology is our tool box offering and we implement their difficult, expensive, or just impossible. In the PC solution to a standard iTAC API interface. With this industry, we saw relatively fast adoption of new approach we can enhance our interface tool box technologies like USB and HDMI, but you’re talking especially for legacy equipment. about a product with a much lower price tag and typical life of 3-5 years. The machines used in the Brain D’Amico: SMT industry are far more expensive and used for much longer, so replacing equipment for Industry Truly seamless integration 4.0 compatible ones will take many years. of data across different manufacturing equipment Thomas Marktscheffel: requires another level of connectivity typically So far, the one-and-only referred to as “Middleware”. industry standard that Rather than “re-invent “does it all” is not available. the wheel”, we at MIRTEC have chosen to Today, most data is still strategically partner with companies like trapped in “silos” – in Cogiscan that specialize in machine-to-machine proprietary formats of communication. These partnerships allow us to various equipment makers, effortlessly connect to virtually any machine within in machines, on the shop floor, and in standalone the manufacturing line without tying up valuable solutions. Furthermore, different parts of the engineering resources. This also overcomes the industry have different histories of using certain hurdle of working with some competitive systems. different standards. I suppose midterm we will see a coexistence of a few standards playing Gerry Padnos: different roles in their industries. Long-term there might be a convergence. However, for the We always try to support time being, openness and flexibility is still key to different protocols and successful communication and integration. Thus, standards, but we have ASM will continue to support industry standards seen past attempts to in use by our customers, responding to customer make a standard fizzle out demand and industry needs. (CAMX). It’s hard to know what will happen with new Explain your strategy to address standards like CFX and Hermes. Because of this, these issues and challenges it’s difficult to justify big investments in software that may not ever be fully adopted. I’m sure James Mok: most equipment manufacturers have the data available for CFX, but making the connection At Dassault Systèmes, we isn’t necessarily driving sales just yet. Rather have an open strategy when than getting distracted from our main mission of it comes to connectivity. We building the highest quality machines, Juki has are not a hardware company partnered with Cogiscan for many years. Since so unlike many other Cogiscan is a neutral third party and already has automation companies that access to the required data for many machine have a strong reliance on manufacturers, it is just one more step for them hardware revenue. This is driving them to adopt to convert the data into CFX format. It’s easier and promote their own standard. We make sure and faster for everyone, which hopefully will help that our software systems are agnostics to all customers reap the benefits faster. types of machines and devices. Also, we have adopted widespread standards of IIoT into our Thomas Marktscheffel: own solutions such as OPC-UA and MQTT, we also work with partners to enhance our ASM is already offering a connectivity such as Cogiscan in the domain coordinated portfolio of of SMT integration, which is at the heart of any products and services for electronic manufacturing nowadays. all integration levels – from circuit board information to Sam Wong: standards and an MES for process data integration There is no single solution in the digital age to an IIoT platform for apps. to meet this diversity. We have answers for partners of choice for all Our strategy is to rely questions regarding data and process integration. on partners to provide We also believe in openness. Connectivity and these interfaces, partners integration are more than technical concepts. No specialize in these domain, supplier can off everything, but he must be willing such as Cogiscan. and able to communicate and cooperate with 14 An insight into Industry 4.0 Sep/Oct 2019

other people and companies in a respectful and connected to a different Interface. If this would trusting manner. We focus on our customers, their be the case, it will at least reduce implementation objectives and their capabilities in everything we efforts for all. do. ASM chose Cogiscan as a partner to support industry standard protocols on top of OIB, Brain D’Amico: because in this cooperation both partners can focus on their core competencies. With Cogiscan While there is certainly a for instance, ASM works closely with a partner trend toward standardization, which has been known for a long time for their we must consider the fact expertise in communication standards. that there are a host of communication protocols What is your perspective relative that are used throughout our to new industry standards such as industry. Proposed standards Hermes, CFX and others you are such as CFX and Hermes are gaining momentum aware of? and will undoubtedly help with M2M communication initiatives. In my estimation, however, there will still James Mok: be the need for middleware in order to bridge the communication gap between standards. The availability of these new standards is making it much The machines used in the easier to connect. Hermes SMT industry are far more provides information of the board in an efficient expensive and used for manner making horizontal much longer so replacing integration across the line equipment for Industry 4.0 seamless. CFX along with OML simplify vertical compatible ones will take integration to enterprise systems and MES. However, the application of such standards many years. depends on the type of individual equipment. For the same reasons mentioned above, we do Gerry Padnos: not see any single standard solution that can meet the needs of smart manufacturing. That is I think the concept of having why we work with a partner like Cogiscan. With easy interoperability is a such a solution, we do not have to worry about great idea and I hope the so many protocols and standards or the burden new standards succeed and of bridging gaps between legacy equipment and really become “standards”. new machines. The connectivity platform solution Writing and approving a that they provide can speak to all the protocol standard is one thing, but and standard in the SMT world. And if ever a getting everyone to use it is another. There seems standard raised to power, we can rest assured to be more excitement for the latest standards than that they would support that too. we have seen in the past. This should be a good thing because if more users and manufacturers get Sam Wong: on board, the adoption will be better and faster. The investments some companies have made in other This is really a good step technology could be a roadblock, but if enough forward and Keysight companies get on board, it will just be a matter of participates in the CFX time for everyone to switch over. and Hermes initiatives. In my view, this is only the Thomas Marktscheffel: beginning and the vast majorities of our test and Getting new standards like inspection customers are coming to speed, CFX and HERMES into the especially on their requirements. Not many market does need some customers had deployed such connectivity and effort and broad support they usually come back to Keysight for advice. We in the market. Midterm, I have also worked with proprietary SMT standards, believe we will continue based on specific customers’ requests. to see a coexistence of a few standards playing an important role in Peter Bollinger: their industries. Long-term there might be a convergence. However, openness, sharing It will take years for a among partners and highest level of flexibility standard like CFX to be is still and will always be key to successful adopted widely. As I have communication and integration. Thus, ASM said, we need to be able will continue to work closely with partners on to connect regardless of all integration levels and support the industry standards or protocols and standards, which are in demand by electronics while these new standards manufacturers and industry needs. might seem to tick a lot of boxes, they are just that new. There are numerous industry standards that are established over decades, and with high levels of adoption in specific markets. Standards are also aging based on technology demands which means at the end that they need to survive at least the life time of machine in production so we get one standard we could use for all machine in production. My hope is that at least the equipment suppliers could agree on the standards that are used for a couple of years so there is no need to have for each of them i40today.com 15

The factory of the future begins with the smart factory of today By Jenny Yuh, Marketing Assistant, Koh Young Technology and Brent Fischthal, Sr. Manager, Americas Marketing, Koh Young America Koh Young’s Intelligent Platform (IP) powers several features within KSMART that help customers analyze and optimize the production process by managing process data from connected SPI and AOI systems. As the preeminent smart factory solutions provider, Koh Young will continue to lead the effort to speed the arrival of the 4th Industrial Revolution. WHAT IS KSMART? anywhere in the network through a highly intuitive user interface for defect detection, real- KSMART is your comprehensive solution for time optimization, enhanced decision making and seamless Smart Factory connectivity and your traceability to improve metrics, increase board portal to Industry 4.0 quality and lower costs by eliminating variance, false calls, and escapes. The KSMART smart factory solution suite, represents a revolution in process optimization From extensive data management offerings to ensure the highest standards of quality and through LM@KSMART, LINK@KSMART, RTM@ reliability on the factory line. KSMART collects KSMART, and in-depth analysis of that data inspection and measurement data from all through SPC@KSMART, to the comprehensive equipment through its KSMART hub to utilize optimization solutions of RMS@KSMART, OPO@ 16 An insight into Industry 4.0 Sep/Oct 2019

KSMART, KSMART integrates 3D solder paste Realizing a smart factory means taking a inspection (SPI) systems and 3D automated practical approach to process and systems optical inspection (AOI) in ways that lead to while examining areas to improve productivity. Industry 4.0 realization. Using real 3D measurement data generated during inspection helps manufacturers define 3D SPI Inspection inefficiencies and boost line efficiency. For example, the LM@KSMART (Library Manager) Why Choose Koh Young module simultaneously deploys programs and KSMART Software? inspection conditions across multiple lines to enhance productivity and data integrity with Koh Young Technology’s KSMART solution consistent performance. Operators can further provides the power of true connectivity to improve line maintenance with other features like pave the way for Industry 4.0 with smart, data- RTM@KSMART (Real Time Monitoring) where by driven analytics the system instantly displays relevant process parameters to remote locations for immediate Koh Young has pioneered true 3D measurement analysis and action. LINK@KSMART provides technology making it the most advanced multipoint views (SPI, Pre-reflow AOI, and Post- equipment of its kind. We have been working reflow AOI) and uses real data management with printer and mounter partners to achieve total and monitoring, so operators can to determine communication and streamline the surface mount actionable insights to optimize processes. line to realize a zero-defect future. Our ability to generate reliable measurement data has given KSMART Software Module Overview rise to our industry-leading, measurement-based KSMART analysis and optimization solutions. LM@KSMART – Library Manager With quality control capabilities and a full lineup of integrated inspection systems, KSMART ties • Deploys optimized programs and conditions to everything together: SPI, AOI and production all Koh Young machines to seamlessly program, machines with AI-powered production analysis for deploy, move, delete, compare, and backup fully-automated control that boosts productivity variant programs for all connected machines. while minimizing costs. Koh Young KSMART Process Flow – Efficient Data Management Focus on Data Management • Offers traceable management of changes via user level identification. Processed data is information, analyzed data • Maximizes productivity by centralizing & is insight, and KSMART’s data management optimizing all programming resources in a modules provide the key to efficient single management system. production, and the highest standards of quality and reliability LINK@KSMART - SPI & AOI Connectivity i40today.com 17

3D SPI Inspection 3D SPI Inspection • Connects all systems within a line to RTM@KSMART – Real-Time Monitoring consolidate inspection results for review to find correlations between processes. • Displays real-time statistics so operators can monitor line performance in real-time. • Streamlines production with regular data- based review, diagnosis and optimization of • Enables a single operator to monitor multiple printing, pick-and-place and reflow processes lines and dozens of machines, which allows via linked inspection results. operators to focus on other critical line tasks. • Provides 3-Point Review including SPI, Pre- • Tracks various machines in the factory, showing AOI and Post-AOI images, trend charts and the best performing machine yield and top inspection results to correlate between five defects list, plus charts reflecting yield, printing, placement and reflow results. availability, X-bar, and sigma all in one page. Koh Young KSMART Process Flow – Smart, Data-driven Analysis Sep/Oct 2019 18 An insight into Industry 4.0

3D AOI Inspection Focus on Data Analysis modules make it easy to maintain, control and upgrade new features via the SPC page. Koh Young delivers an innovative Smart Factory AI platform for fully-automated • Supplies a straightforward visualization of real- process optimization to harness the power of time production and results for all lines. true connectivity • Offers charts and graphs relating to refined The power of KSMART solution lies in its analytical analysis results from production, including power. SPC@KSMART (Statistical Process configurable charts for user-specific parameters. Control) provides a straightforward visualization of real-time production and inspection results, • Drills down to the component level to identify including configurable charts for user-specific the root cause to minimize false calls or tighten parameters. Users can identify the exact defect tolerance levels to prevent future escapes, as origin by checking false calls and NG parts from well as evaluate and optimize default settings. the dashboard, as well as evaluate and optimize default settings by comparing actual results with Focus on Process Optimization average, minimum, and maximum values. For example, if the process was stable, operators can KSMART delivers the opportunity to tighten the tolerance to prevent escapes. finally realize a fully automated and advanced manufacturing facility through a comprehensive infrastructure for autonomous process optimization Koh Young KSMART Process Flow – Autonomous Process Optimization KSMART Software Module Overview Harnessing the power of the Koh Young IP (Intelligent Platform), OPO@KSMART (Offline SPC@KSMART – Statistical Process Control Programming Optimizer) provides an intuitive graphical simulation tool to review identified • Provides a cross-platform multi-line process defects with accumulated historical real data from analysis application that includes production all lines, which avoids unnecessary downtime. performance analysis by product. KSMART reliably allows users to foresee the impact of fine-tuning without stopping the line. • Delivers detailed inspection results with visual Moving forward, KSMART autonomously renders graphics and accumulated statistics through complex process optimization decisions typically deep inspection data analysis tools, including reserved for dedicated process engineers. In charts, histograms, and more. short, it is a quantum leap towards the realization of a Smart Factory. We designed the modular • Instantly analyzes data visualized with relevant platform for future growth and expansion, so indicators to compare board performance and when Koh Young releases new modules or when identify process deviations. the process requires additional smart factory • Comprehensive and extensible cross-platform i40today.com 19

modules, the manufacturer can simply implement Koh Young has developed the network tools the upgrades easily anytime. to connect with multiple suppliers and simplify communication across the entire PCBA line. KSMART Software Module Overview As the absolute market and technology leader, RMS@KSMART – Remote Monitoring System Koh Young is using its IP (Intelligent Platform) to achieve its vision with a focus on next-generation • Removes defect reviewers from the production cooperative efforts that expand process floor while providing real-time remote access to capabilities and factory performance. To this end, each piece of equipment throughout the network. the company has established three additional R&D centers worldwide to facilitate a quantum leap in • Provides direct control from a web browser technological leadership and competitiveness. at a remote location to review all data, Koh Young can apply the IP (Intelligent Platform) machine status indications including yield and to its current areas of expertise, while paving the availability, and systems. way for new markets and industries beyond SMT. OPO@KSMART – Offline Program The gap between the digital and physical world Optimizer is closing fast, replaced by an entirely new way of manufacturing. • Allows operators to load and debug identified defects and enables measurement- About Koh Young Technology based finetuning through simulations with accumulated historical data from all lines. Koh Young is the leading 3D measurement-based inspection equipment and solutions provider, • Simulates the results of any adjustments performs an essential role for quality control without affecting production and deploys new and process optimization across a growing settings to all machines in the production line. set of industries including printed circuit board assembly, machining and assembly process Creating Smarter Factories Together manufacturing, semiconductor manufacturing, and various medical fields. In addition to its corporate Koh Young is leading the effort to harness the headquarters in Seoul, Koh Young has sales and power of connectivity and create a smart factory. support offices in Germany, Japan, Singapore, For instance, its Koh Young Process Optimizer China, Mexico, and the United States. These local (KPO) includes four interlinking software modules facilities ensure it sustains a close relationship with that exercise complex algorithms to develop its growing customer base, while providing them closed-loop process recommendations. The with access to a global network of inspection and Machine-to- Machine (M2M) connectivity drives measurement experts. Contact us today to see the smart factory vision one step further by how Koh Young can help your business. enabling automatic SMT line maintenance. Finally, working with its printer and mounter partners, www.kohyoung.com 20 An insight into Industry 4.0 Sep/Oct 2019

Current partners Koh Young America, Inc. Find us at 1950 Evergreen Blvd., Suite 200, Duluth, GA 30096 +1-470-374-9254 | [email protected] BOOTH #609 kohyoung.com i40today.com 21

Yamaha delivers configurable modular transfer for robot assembly cells By Oumayma Grad, Marketing Communications Manager, Yamaha Motor Europe N.V. Flexible LCM100 linear conveyor module boosts efficiency over conventional belt- and-roller transfer The LCM100 linear conveyor module, a unique drives include low friction and lag-free solution for robotic assembly by Yamaha Factory acceleration and deceleration. Automation Section, enables cleaner, quieter, and more flexible workpiece transport. By The robust and stable sliders, combined with simplifying production startup, reducing cycle positioning accuracy better than ± 0.015mm, times, and increasing accuracy, the modules allow assembly tasks to be performed directly boost productivity and yield, while also cutting on the module. With no need to de-palletise and noise, increasing reliability, and reducing the size move workpieces from a conventional conveyor of robot assembly cells. to a work-table, tact time can be faster and assembly cells smaller and more cost-effective. Unlike conventional belt-and roller conveyors, Without physical microswitches or end stops, each LCM100 module contains an independently slider stop positions can be quickly and easily controlled high-speed linear motor that enables reprogrammed in software. bidirectional movement and programmable speed up to 3000mm/second. Production lines In addition, LCM100 modules are featured for can be configured optimally, not dependent on easy interchangeability and reconfiguration, enabling assembly cells to be setup quickly and the longest process cycle time, and utilize cost-effectively, especially for short production inline equipment more efficiently runs. Modules can also be removed or replaced to save space and capital easily to minimise downtime, and reused in other expenditure. Further assembly cells to maximise utilisation. benefits of the linear The modules are available in standard 480mm and 640mm lengths and carry 15kg maximum payload. A 400mm circulation module is also available. www.yamaha-motor-im.eu 22 An insight into Industry 4.0 Sep/Oct 2019

SAN JOSE [email protected] Join our Industry 4.0 Event San Jose, October 2-3, 2019 Fremont Marriott Silicon Valley Hotel REGISTER NOW i40connectforum.com i40today.com 23

“Industrial AI Company” PerfectPattern expands AI competence By i4.0 Today Editor PerfectPattern, which develops innovative artificial intelligence solutions primarily for the manufacturing industry, is continuing to expand its competence in the field of AI. Michael Klaput, who was responsible for the AI-based valuation of derivatives as Vice President Quantitative Analyst at Barclays Bank in London, joined PerfectPattern’s AI development team as Principal Researcher in July. such as mastering a large number of often- contradictory goals like cost and time savings, capacity increases and more. Specialists in the fields of artificial intelligence and machine learning are very much in demand AI expert Michael Klaput joins the AI Study friends reunited development team of the “Industrial AI Company” PerfectPattern as Principal Researcher Michael Klaput joins PerfectPattern’s KI development team headed by CTO Christian PerfectPattern develops AI solutions and Paleani. Both already know each other from their technologies especially for manufacturing studies at the Technical University of Munich. companies that maximize the potential of artificial Klaput studied physics with a focus on theoretical intelligence. Its aim is to monitor and control physics and mathematical methods. His diploma industrial processes so that companies can thesis dealt with string theory and he ultimately operate more reliably and efficiently and can earned a PhD in theoretical physics from the react flexibly to changing requirements. University of Oxford. Solving such highly complex mathematical Klaput’s first career steps were at the Bank of problems requires intricate computations, America Merrill Lynch and Barclays Bank in London, where he used mathematical methods to optimize the valuation of derivatives. Mathematics is the language of reality “I basically want to understand reality better, and as a theoretical physicist I have learned to make reality understandable with the help of mathematics. Through mathematics as a language, I start a dialogue with reality, so to speak. This not only includes logical thinking, but also intuition, fantasy and creativity,” Klaput explained. “It is crucial to include coincidences with mathematical precision. This has already been successfully applied in the financial world in recent years. With PerfectPattern, I can now apply this expertise to industrial production environments.” AI and ML specialists are in demand “Specialists in the fields of artificial intelligence and machine learning are very much in demand — all the more reason for us to be pleased that 24 An insight into Industry 4.0 Sep/Oct 2019

PYTHIA Others 1 Unsupervised data synchronization 1 Data synchronization by data scientists 2 Unsupervised Anomaly Detection for Data Cleaning 2 Data Cleaning by data scientist 3 Unsupervised extraction of relationships including 3 Extraction of relationships including response times by data scientist and response times domain expert 4 Automated creation of model of time series 4 Root Cause Analysis and creation of model of time series by data scientist 5 Prediction of target quantity based on delay SDEs 5 Prediction of target quantity by data scientist 6 Learning of impact on expectation of target quantity 6 Creation of model to affect root causes by data scientist PerfectPattern´s PYTHIA is a product for unsupervised and automated regression analysis, anomaly detection, classification and time series prediction on real-time data streams. This table compares the methodology of typical data analysis (others) step by step with PYTHIA. Amount of data PYTHIA ANN needed Small Huge Training Global Local Calculation Fast Fast speed No Tends to No A lot Overfitting Full Not clear why it works Tweaking Transparency PerfectPattern´s PYTHIA provides a new way of pattern recognition. Within seconds, PerfectPattern´s KI based sPrint One calculates a This table compares pattern recognition of PYTHIA with artificial complete and cost-optimized print plan from unlimited pools of unsorted neuronal networks (ANN). print orders. This not only yields significant savings, but also enables dynamic print planning. Print businesses any size can exploit significant savings with PerfectPattern´s KI based sPrint One technology. we have been able to recruit Michael, a proven AI products and technologies for CORTEX is a decision-making technology that expert in these areas,” says Fabian Rüchardt, manufacturing industries makes decisions based on global objective CEO of PerfectPattern. “Christian and Michael functions using reinforcement learning. have known one another for many years from With sPrint One and Kayros, PerfectPattern CORTEX’s special capabilities lie in the fact their studies in Munich. This duo will be hard to already offers two AI products for manufacturing that it can solve non-linear decision problems beat when it comes to innovation and creativity companies. sPrint One is used for dynamic print in a very short time, with a very large number of in the application of mathematical expertise.” planning and the calculation of printing forms, conflicting influencing parameters. including so-called gang forms, in digital and Unasked questions, unexplored answers offset printing. Kayros plans any job pool on Well connected in international industry very complex production networks with the aim Klaput explained his decision to join of minimizing make-ready costs and meeting PerfectPattern maintains partnerships with PerfectPattern, saying, “After more than five production deadlines. well-known companies worldwide, including years in finance, I wanted to deal with new topics Cimpress, Kodak, Sappi (OctoBoost) and and problems. I am fascinated by the still young These products are based on two technologies Voith. In addition, PerfectPattern has recently topic of digitization, with so many unasked developed by PerfectPattern. PYTHIA is a established a joint subsidiary with MPDV, a questions and unexplored answers. In addition, platform product for pattern recognition, time leading provider of manufacturing IT. The goal I appreciate the culture at PerfectPattern. The series prediction and anomaly detection in is to develop and provide software components development team is protected from routine real-time data streams. By combining methods for artificial intelligence in manufacturing. tasks and has the necessary leeway.” from deep learning, stochastic calculation, infinite dimensional geometry and quantum www.perfectpattern.de/en field theory, it independently uncovers even the most hidden patterns. i40today.com 25

Q&A with Drew Elhers, Global Director, Portfolio Marketing Enterprise Software, Zebra Technologies By i4.0 Today Editor As the digital revolution permeates and the amount of data surges, unlocking the power from that data has become ever more critical. Extracting actionable insights from this vast ocean of information enables enterprises to transform digitally, drive better business performance, streamline workflows, identify supply chain bottlenecks and predict better business outcomes. Taking the heavy lifting out of data sifting, Zebra’s Business innovators across retail, healthcare, Savanna Data Services facilitates this process transportation & logistics and manufacturing as and digs deep into big data for actionable well as software developers involved in printing, business insights. With its cloud-based capability, scanning, barcoding, RFID and indoor location Zebra’s data intelligence platform is designed to tracking solutions will have access to vital APIs, assist software developers, ISVs and business blockchain traceability, barcode intelligence innovators in building cutting-edge applications and recall notifications. All hosted on Zebra’s that unlock the potential of data from Zebra Developer Portal, Savanna Data Services will be devices, without the expense of maintaining sold via a web-based, self-service model and can and managing a platform. Drew Ehlers, Global be accessed 24/7. Director, Enterprise Software, Office of the CTO from Zebra Technologies has the answers to Early adopters of Savanna Data Services include your question and explains the potential for Doddle, StayLinked and Qodenext: businesses everywhere. Doddle Q. What is Savanna Data Services? Doddle is one of UK’s largest providers of click A. Savanna Data Services is a cloud-based & collect services, the fastest growing fulfilment platform providing a more complete business method in the UK. Savanna has provided Doddle picture to help transform operations. This with the tools to quickly replicate its click and new capability is part of our breakthrough collect solutions to meet retail market demands. data intelligence platform – known as Zebra As a simple use case, Doddle has used a ‘print- Savanna – which allows adopters to collect and from-cloud’ API to reduce the deployment time of process data from Zebra hardware in real time. It its ‘ship from store’ proposition. Savanna is also enables business managers to turn raw data into a key component of Doddle’s returns platform actionable insights that digitally transform the – using the blockchain ledger to distribute data way businesses drive performance, streamline across the supply chain to enable intelligent workflows, identify supply chain bottlenecks and decisions – reducing handling costs and getting predict better business outcomes. items back on sale more quickly. Savanna Data Services provides real time visibility into the 3rd The platform’s modus operandi unlocks the party systems to update inventory and refines potential of data to build more intelligent, productive work-flow efficiency. applications and solutions for digital transformation. StayLinked Q. What challenges does Zebra Savanna Data Services address? For over 20 years, StayLinked has been delivering world-class solutions for mission- A. Zebra designed Savanna Data Services’ to critical, host-based systems in the warehousing, unify API’s (Application Programming Interface) manufacturing, transportation, and retail and developer tools and to empower business industries. StayLinked experts have recently innovators to easily and quickly build secure, introduced a new paradigm in terminal emulation scalable digital services. The enriched API and session management for mobile devices. platform layer enables developers to innovate, StayLinked uses the Savanna platform to enhance design and rapidly deploy customised, cutting- its ability to deliver customers a complete view edge solutions without the expense of maintaining of their operations, integrating a solution into and managing a platform. the existing Wi-Fi systems by providing relative positioning for workers from an aggregated top- 26 An insight into Industry 4.0 Sep/Oct 2019

i40today.com 27

down view of the work site, optimising workflow. It Location Engine: Savanna takes the location extracts valuable Zebra device information as well data collected by Zebra MotionWorks™ and as complementing and enhancing StayLinked iQ translates it into actionable information specific to data, helping its customers accurately measure a customers’ operational context. Using proven everything that impacts productivity, giving them algorithms and rules engines, it uses the tracking a competitive edge and minimizing costs. of assets, goods and people to generate impactful and actionable insights that can transform the Qodenext way companies do business. QodeNext is a leading supply chain, traceability These insights can then be accessed and used technology, consulting and service company, to power new applications and solutions and providing a single point of contact for an developers can easily access Savanna APIs via a enterprise’s traceability needs. QodeNext self-service portal (24x7). selected Savanna’s print API and blockchain authentication engine to enable total transparency Q. Why is this service significant? for their customers needing accurate visibility of the raw materials used during the production and A. The Savanna platform and the new capabilities distribution process. This automation has resulted now offered with Savanna Data Services are the in significant inventory management savings and building blocks for digital transformation with APIs increased productivity and staff optimisation. for Internet of Things (IoT), artificial intelligence and machine learning applications. Agility and Encompassing building blocks such as IoT, flexibility are key aspects of decision making and artificial intelligence and machine learning, the Savanna provides a refined system to extract the platform powers new applications and solutions, correct data and deliver actionable insights for allowing them to transform various data points business success. into actionable insights. Using existing APIs will empower business Q. What are the key features of Savanna innovators and developers to rapidly build cutting- Data Services? edge solutions, shorten development cycles and reduce costs. Adopters of the Savanna Data A. There are several features offered by Savanna Services can take advantage of the building blocks Data services which enable enterprises to realise available to them, such as artificial intelligence and business aims through data driven actionable machine learning, to develop their own fluid, agile, insights. Firstly, Savanna Data Services offer bespoke applications and solutions. new cloud-accessible APIs for ISV (Independent Software Vendor), software developers and Overall, combining the real-time data that Zebra business innovators to build their own solutions. devices capture on the shop floor, with the The three main features of Savanna are: operational data captured by Savanna – including machine learning and prescriptive analytics - Data services: Savanna collects device-level business managers will have the insights they data and makes it accessible via APIs to increase need to make better, faster, smarter decisions. workflow efficiency and productivity and can integrate with existing hardware, software and Q. How do I access the services? data ecosystems to provide a more complete business picture. This is used to improve current A. Savanna Data Services APIs is sold directly applications and create new solutions. through Zebra’s self-service Developer Portal to software developers, Independent Software Data and analytics platform: Savanna aggregates Vendors and business innovators from retail, and analyses data from Zebra devices and transport and logistics, manufacturing and intelligent infrastructure, combining real-time data healthcare sectors. New APIs will be released with historic data, to create a complete picture on a rolling basis, and “Sandbox” APIs will be and intelligent insight for customers. Using data available to customers and developers for testing collection, analytics, artificial intelligence and prior to release. machine learning, Savanna powers next-generation applications and solutions to create data-powered www.zebra.com environments and guide real-time actions. 28 An insight into Industry 4.0 Sep/Oct 2019

Digitalization powers electronics innovation. In an industry that demands new products at an unprecedented rate, electronics companies are increasingly relying on “smart manufacturing” to address the challenges of complexity, customization, compliance, globalization and customer expectations. That’s where we come in. Our Digital Enterprise solutions help optimize your business so you’re better equipped to initiate and respond to disruptive innovation. siemens.com/plm/electronics i40today.com 29

Making machine- learning inference meet real-world performance demands By Daniel Eaton, Sr Manager, Strategic Marketing Development, Xilinx FPGAs offer the configurability needed for real-time machine-learning inference, with the flexibility to adapt to future workloads. Making these advantages accessible to data-scientists and developers calls for tools that are both comprehensive and easy to use. Real-time Machine-Learning Inference efficiently in and out. This requires features such as a flexible memory hierarchy and adaptable Machine learning is the force behind new services high-bandwidth interconnects. that leverage natural voice interaction and image recognition to deliver seamless social media or Contrasting with these demands, the GPU- call-center experiences. Moreover, with their ability based engines typically used for training neural to identify patterns or exceptions in vast quantities networks – which takes time and many teraFLOPS of data related to large numbers of variables, of compute cycles – have rigid interconnect trained deep-learning neural networks are also structures and memory hierarchy that are not transforming the way we go about scientific well suited to real-time inference. Problems such research, financial planning, running smart-cities, as data replication, cache misses, and blocking programming industrial robots, and delivering commonly occur. A more flexible and scalable digital business transformation through services architecture is needed to achieve satisfactory such as digital twin and predictive maintenance. inferencing performance. Whether the trained networks are deployed for Leading Projects Leverage Configurability inference in the Cloud or in embedded systems at the network edge, most users’ expectations Field Programmable Gate Arrays (FPGAs) that call for deterministic throughput and low integrate optimized compute tiles, distributed latency. Achieving both simultaneously, within local memory, and adaptable, non-blocking practicable size and power constraints, requires shared interconnects can overcome the an efficient, massively parallel compute engine traditional limitations to ensure deterministic at the heart of a system architected to move data throughput and low latency. Indeed, as machine- learning workloads become more demanding, cutting-edge machine-learning projects such as Microsoft’s Project BrainWave are using FPGAs to execute real-time calculations cost- effectively and with extremely low latency that has proved to be unachievable using GPUs. Another advanced machine- learning project, by global compute-services provider Alibaba Cloud, chose FPGAs as the foundation to build a Deep Learning Processor (DLP) for image recognition and analysis. FPGAs enabled the DLP to achieve simultaneous low latency and high performance that the company’s Infrastructure Service Group believes could not have been realized using GPUs. Figure 1 shows results Figure 1. Alibaba Cloud DLP performance and latency comparison. from the team’s analysis with 30 An insight into Industry 4.0 Sep/Oct 2019

a ResNet-18 deep residual network that shows tools, and a runtime, collectively called xfDNN, how the FPGA-based DLP achieves latency of which ensure the neural network delivers the just 0.174 seconds: 86% faster than a comparable best possible performance in FPGA silicon. GPU case. Throughput measured in Queries Per Second (QPS) is more than seven times higher. The ecosystem also leverages Xilinx’s acquisition of DeePhi Technology by utilizing the DeePhi Projects such as Microsoft’s BrainWave and pruner to remove near-zero weights and Alibaba’s DLP have successfully established new compress and simplify network layers. The hardware architectures capable of accelerating DeePhi pruner has been shown to increase neural AI workloads. This is just the beginning of the network speed by a factor of 10 and significantly journey that will ultimately make machine-learning reduce system power consumption without acceleration widely available to Cloud-services harming overall performance and accuracy. customers, as well as industrial users and the automotive community who are more often Whether the trained seeking to deploy machine-learning inference in networks are deployed embedded systems at the network edge. for inference in the On the other hand, some service providers are Cloud or in embedded keen to infuse machine learning into existing systems at the network systems to enhance and accelerate established use cases. Examples include network security, edge, most users’ where machine learning enhances pattern expectations call for recognition to drive high-speed detection of deterministic throughput malware and dangerous exceptions. Other opportunities include using machine-learning and low latency. applications such as facial recognition or disturbance detection to help smart cities run When it comes to deploying the converted more smoothly. neural network, ML-Suite provides xDNN custom processor overlays that abstract designers from the AI Acceleration for Non-FPGA Experts complexities of FPGA design and utilize the on-chip resources efficiently. Each overlay typically comes Xilinx has established an ecosystem of resources with its own optimized instruction set for running that let users take advantage of the potential of FPGAs to accelerate AI workloads in the Cloud or at the edge. Among the tools available, ML-Suite (figure 2) takes care of compiling the neural network to run in Xilinx FPGA hardware. It can work with neural networks generated by common machine- learning frameworks including TensorFlow, Caffe, MxNet, and others. A Python API makes interacting with the ML-Suite easy. Figure 2. Xilinx ML-Suite provides an ecosystem of resources for machine-learning development. Because machine-learning frameworks tend various types of neural networks. Users can interact to generate neural networks based on 32-bit with the neural network via RESTful APIs, while floating-point arithmetic, ML-Suite contains working within their preferred environment. a quantizer tool that converts it to a fixed- point equivalent that is better suited to being For on-premises deployments, Xilinx Alveo™ implemented in an FPGA. The quantizer is part of accelerator cards remove hardware development a set of middleware, compilation and optimization challenges and simplify infusing machine-learning with existing applications in the data center. i40today.com 31

The ecosystem supports machine-learning At present, commercial machine-learning deployment in embedded or edge use cases applications tend to be focused on image handling leveraging not only the pruner but also a quantizer, and object or feature recognition, which are best compiler, and runtime from DeePhi Technology handled using convolutional neural networks. This to create high-performing and efficient neural could change in the future as developers leverage networks suitable for resource-constrained the power of machine learning to accelerate embedded hardware (figure 3). Turnkey hardware tasks such as sorting through strings or analyzing such as the Zynq™ UltraScale™ 9 card and Zynq unconnected data. Workloads like these are better 7020 System-on-Module simplify hardware served by other types of neural networks such as development and accelerate software integration. random forest or Long Short-Term Memory (LSTM) networks. If the hardware must be updated to Figure 3. ML-Suite provides tools optimized for Cloud and edge/embedded machine-learning. host different types of neural networks needed to There are also a number of innovative independent ensure fast compute times with low latency, this software vendors who have built CNN inference could take months or years. overlays that can be deployed to FPGAs. Mipsology has built Zebra, a CNN inference Building an inference engine based on processors accelerator that can easily replace CPU or GPU such as GPUs or custom ASICs, which have a fixed and supports a number of standard networks (ie architecture, leaves no easy or fast way to update Resnet50, InceptionV3, Caffenet) as well as custom the hardware. The pace of AI development is frameworks and has demonstrated amazing currently outstripping that of silicon, so a custom throughput at lowest latency, such as Resnet50 at ASIC that may represent the state of the art at the 3,700 images/second. beginning of its development will be outdated Omnitek DPU is another example of an inference even before it is ready to deploy. overlay that runs very high performance DNNs on an FPGA. For example, on GoogLeNet Inception-v1 In contrast, the reconfigurability of FPGAs and the CNN, the Omnitek DPU performs inference on sheer flexibility to customize the resources are key 224×224 images using 8-bit integer processing strengths that enable these devices to keep pace at over 5,300 inferences per second on a Xilinx with the evolution of this exciting field. We already Alveo Data center accelerator card. know that FPGAs are well suited to low-latency Reconfigurable Compute for Future Flexibility clustering used for unsupervised learning, which In addition to the challenges associated with is another emerging branch of AI and particularly ensuring the required inferencing performance, well suited to tasks such as statistical analysis. developers deploying machine learning must also bear in mind that the entire technological landscape Using a tool such as ML-Suite to optimize and around machine learning and artificial intelligence compile the network for FPGA deployment allows is changing rapidly; today’s state-of-the-art neural developers to work at a high level in their own networks could be quickly superseded by newer, environment without needing FPGA expertise to faster networks that may not fit well with legacy direct the compiler’s decisions, while retaining the hardware architectures. flexibility to reconfigure the hardware in the future to support later generations of neural networks. Conclusion FPGAs are known to provide the performance acceleration and future flexibility that machine- learning practitioners need; not only to build high- performing and efficient inference engines for immediate deployment, but also to adapt with the rapid changes in both the technology and market demands for machine learning. The challenge is to make the architectural advantages of FPGAs accessible to machine-learning specialists and at the same time help ensure the best performing and most efficient implementation. Xilinx’s ecosystem has combined state-of-the-art FPGA tools with convenient APIs to let developers take full advantage of the silicon without having to learn the finer points of FPGA design. www.xilinx.com 32 An insight into Industry 4.0 Sep/Oct 2019

Collaborate 12-15 Nov 2019, Messe München Halle 4, Stand 466 Advise Formulate Invent Partner Test Research Tailored Solutions Refine Manufacture Distribute Approve Our winning Comply formula for superior Support Globally electro-chemicals When every connection counts, you can count on Electrolube’s electro-chemical expertise. With a 77-year pedigree, a growing presence in 55 countries and production in 3 global locations, we have the products, research facilities, resources and personal expertise to engineer solutions to your manufacturing challenges. Make contact and discover what makes Electrolube the solutions people for leading manufacturers worldwide. +44 (0)1530 419600 www.electrolube.com Electronic & General Conformal Encapsulation Thermal Management Contact Maintenance Purpose Cleaning Coatings Resins Solutions Lubricants & Service Aids i40today.com 33

8 signs your material management needs an upgrade By Radu Diaconescu, VP, Business Development, Swissmic The heart of the electronics factory is in its machines. However, even the lines and factories with the smartest and most versatile machines rely on intelligent and consistent support received by the central processes to reach their maximum capability. Material Management is one of the most critical related processes on the factory floor. Such smart processes for the success of an electronics tools deliver complete transparency from incoming factory. In the age of the 4th industrial revolution, materials to placement information and let you many different components of increasingly see where materials are located at all times. This smaller quantities must be procured, stored and gain in transparency reduces operator trips and then transported on the factory floor to wherever material travel as components can be supplied they are needed. The materials cost of goods sold where needed ahead of time. Any changes in (MCOGs) is between 75 percent and 85 percent order sequence is communicated in real time to for the average electronics product. There’s the lines themselves, smart shelves, warehousing, no doubt best practices in procurement and etc. This information then makes it possible for an tight inventory turns helps electronics hardware MES system to “order” just in time replenishment companies compete in the marketplace, material. The result is a seamless material flow on especially with launching non-unique devices the factory floor. A smart and integrated material where price and features/functionality drive management system generates tangible results purchase decisions over brand awareness. by making the production agile and reducing the operations complexity. In the environment of an electronics factory, the availability or lack of parts directly impacts what If you see one of the following phenomena jobs can be run. Additionally, the rate of production happening on your factory floor, this means that and related consumption of parts directly your material management desperately needs to impact the production plan and aid in deciding a be upgraded to Industry 4.0 standards. management policy such as overtime or time off. Manufacturing productivity can be increased if the Lack of visibility regarding materials in storage supply of material runs smoothly and there is no and on the factory floor shortage of material disruption. If there isn’t enough transparency regarding There’s no doubt the quantities and the location of material reels best practices in in a factory, this can lead to excessive buffers procurement and tight of materials. inventory turns helps electronics hardware Mistakes during material issue/provisioning companies compete in the marketplace... There’s nothing worse than picking the wrong part number. This can ruin a whole production batch and This is not an easy task. Safety stocks, extended results in significant costs associated with scrap of searching, material movement and material- products, and rework costs as a best scenario. related line stops are symptoms of weaknesses and missing tools in SMT-specific material Inability to locate materials/partial quantities in management. The solution lies in integrated setup preparation areas, on changeover tables, material flow solutions that don’t just administer lines, etc. inventories, but actively support all material- If the material movements are not properly recorded and tracked, this can lead to a huge increase in the time it takes to locate the needed materials, or more significantly, missed customer deliveries. Material-related line stops The delivery of material to the production line is too slow and inefficient and the production lines have to stop so that a certain reel to be found in the warehouse. Manual processing of printed order lists/bills of material 34 An insight into Industry 4.0 Sep/Oct 2019

The lack of a standardized material labels leads Start your journey to manual entry of parameters which in turn towards Industry increases the incoming material processing time 4.0 with the and decreases the efficiency. Swissmic Smart Factory Equipment Placement machines not integrated, leading to major discrepancies between theoretical and actual stocks If the production line is not properly integrated with the material management system, the actual consumption rates are not fed back into the system, therefore the actual level of stock of a certain part number is never known. A smart and integrated material management system generates tangible results by making the production agile and reducing the operations complexity. Line staff must request fresh supplies manually Without an automated system, line operators are continuously watching material levels, finding the appropriate stock location and retrieving or requesting the materials. An automated system that monitors these levels and programmatically ‘orders’ them also eliminates the possibility of a part running out before being noticed (stopping your production) and also removes the time associated with these manual tasks – also freeing your operators to focus on improving machine performance and related process changes that will ultimately improve your production. By using the Swissmic Agilink Smart Shelves, together with the Aegis’ FactoryLogix MES, your factory floor material flow becomes fully traceable, transparent and integrated with your production line. The benefits of using this bundled solution are: • Real time material management analytics. What materials are about to be needed on the SMT machines, what is the actual consumption rate of certain materials, is there enough stock of a certain material for the next job? These questions won’t bother you again. • Full traceability of materials on the factory floor. Seamlessly transfer the materials from the warehouse Agilink Smart Shelves to the Agilink Smart Mobile shelves and finally to the assembly line, without losing track of any movement. • Prevent errors and increase efficiency. Your employees will never pick up the wrong material reel ever again. More so, by scanning a certain BoM or work order, they will be able to pick up the exact materials needed, without wasting time searching for the needle in the haystack. The Agilink Smart Shelf is the latest Industry 4.0 solution that combines the latest advances in IoT, AI and sensor technology to deliver a seamless material management solution for the smart electronics factory. www.swissmic.com i40today.com 35

Minimising the risk of cyber breaches in the manufacturing industry By Paul Robinson, Management Consultant within WWT’s Global Digital Transformation Group Manufacturers across the industry are embracing advances in Industrial Internet of Things (IIoT) technologies, which utilise sensor-based data to automate processes, optimise resources and maximise profitability. The operational technology (OT) systems in Creating cultural change these industrial environments were traditionally closed off, or ‘air gapped’, from core IT systems. The path to robust security lies in collaboration But for the first time they are getting connected and alignment of teams across the digital factory. to the outside world through plant networks, IT and OT departments must start to focus on connected Industrial Control Systems (ICS) the same business objectives and work to the and Supervisory Control and Data Acquisition same operating models. For example, businesses (SCADA) environments. should look to ensure the OT team’s focus on uptime and operational safety is combined with While connecting previously isolated industrial the cybersecurity routines familiar to IT, such as systems to the wider IT network and the internet patching systems and maintaining firewalls and anti- provides tremendous value for companies, it also virus software. exposes them to a host of dangerous security risks. The responsibility of team alignment extends to New systems, new threats an executive level. Understanding and addressing operational risks requires many stakeholders, Many organisations’ OT systems have been in from boardroom to IT to manufacturing. IT must operation for years without being subject to the quickly learn the business of manufacturing, and same upgrade and replacement cycles as their manufacturing must quickly learn the risks they IT systems. This can expose vulnerabilities that, face and how technology can be safely applied to if exploited, may have significant implications address those risks. for both the organisation’s network and for the connected production systems. By working closely together towards the same objectives, staff can collaborate to find the right As industrial systems become more technical and cultural solutions to new security connected, the risk of network breach can threats that do not create a barrier to innovation. have significant – potentially catastrophic – consequences. According to Fortinet, nearly 6 Securing the digital factory in 10 organisations using SCADA or ICS have experienced a breach in those systems in the Due to the hazardous nature of compromised past year. machinery and physical infrastructure, the risks organisations face are far greater in an industrial There have already been several cases that context than those posed in other industries. show the potential impact of an industrial cyber It’s therefore the responsibility of manufacturing breach. For example, the NotPetya attack in executives to look at every element of their critical 2017 caused a global meltdown resulting in systems to understand how, when and where to more than $10 billion in damage. As well as implement stronger security controls. Crucially, this affecting enterprise IT systems, this attack must include real-time unknown threat identification wreaked havoc on industrial systems. Nordic of internal, external and third parties. Only then can shipping company A.P. Møller-Maersk had its risk assessment and remediation extend through entire global network brought to a standstill by OT Cyber expertise, technology and tooling. the attack. This company alone experienced damage amounting to between $250 million Globally-oriented attacks of the scale of NotPetya and $300 million. are when, not if, scenarios. Decision makers must work to align OT and IT teams, to unlock the The potential threat is not limited to the logical potential of IIoT without becoming an open target realm. In Germany, a cyber attack on a steel mill for cyber crime. resulted both in the locking of its IT systems, and in physical damage. Control systems were www.wwt.com manipulated to such a degree that a blast furnace could not be properly shut down, resulting in destruction of equipment. 36 An insight into Industry 4.0 Sep/Oct 2019

SMART MANUFACTURING FOR ELECTRONICS Valor® Manufacturing Analytics is designed specifically for the PCB Manufacturing industry, providing executives, line managers and manufacturing engineers with crucial information needed to deliver quality products on time. For more information call 1-800-547-3000 or visit mentor.com/valor i40today.com 37

IoT and pollution: A breath of fresh air By Nick Sacke, Head of IoT and Products, Comms365 In light of the recent debate surrounding climate change, the topic of air pollution is increasingly becoming a major concern for many cities around the world. And with research from the World Health Organisation revealing that 91% of the world’s population live in places where air quality exceeds WHO guidelines, cleaning up these pollutants is becoming even more challenging. Although Europe saw a decrease in emissions of air pollutants by more than 2.5% in 2018, concentration still remains high. It is therefore necessary to measure air quality and keep it under control; something the Internet of Things (IoT) is already helping with. With a disproportionate level of pollution actionable results. For example, in the city of permeating cities in relation to inhabitant numbers, Uppsala, Sweden, the GreenIoT project is creating smart cities, which are built on the mantra of an integrated solution for an environmental sustainability, are actively fighting the battle against sensing system by implementing real-time air air pollution. Making moves towards a smarter pollution monitoring through wireless sensors on infrastructure or simple solutions such as traffic public transportation vehicles. Through this sensor monitoring to help tackle the issue, many cities data, which is made available for governmental are implementing green legislation and creating agencies, they are able to control traffic and as greener spaces. For example, Paris is expanding a result, make informed city planning decisions, their car-free zones and Tokyo is investing further such as rerouting traffic away from highly polluted into renewable power, becoming progressively areas. Another way that city centre traffic can be more sustainable and resilient – and tangible reduced is through smart parking. IoT enabled benefits are already being realised. parking, which is able to identify empty car spaces, is able to not only decrease the amount of Co2 Although highly polluted cities such as Delhi, India emissions, but also save drivers time and money and Beijing are utilising smart sensors in order to through increased traffic flow, promoting more alert residents when air pollution levels are high, sustainable urban mobility. this is not a long term solution to combat the source of the problem. IoT opens up better pollution data Recent Research from the British Lung Association than ever before, which can be used to advance revealed that 248 hospitals and 2,220 GP 38 An insight into Industry 4.0 Sep/Oct 2019

practices are located in areas where air pollution is partnered IoT driven air quality solutions, becomes significantly above the World Health Organisations an increasingly valuable solution to undertake limit for fine particulate matter (PM2.5). With pollution. Through established sensors, councils hospitals in polluted areas proven to contribute to have the ability to help reduce pollution through worse outcomes for vulnerable patients who are changing transport routes and urban planning. more susceptible to the harmful effects, citizens Advances in technology have also led to the are left exposed, increasing concern for public availability of the Low Power Wide Area Network health. Britain has already pledged legislation (LPWAN) sensors as a low cost alternative to to address pollution in London hospitals with fixed monitoring systems, which makes them an monitors to measure toxic air levels planning to be increasingly attractive option. implemented this year, however the government still has a long way to go, especially when it Many cities have already taken the first step comes to building a framework for post- Brexit towards integrating IoT solutions into their environmental law. ecosystem, and as a result are realising the long- term benefits. With the help of technology vendors Reducing pollution is paramount to the future of in alliance with the government and conviction smart cities, and it doesn’t require a plethora of from the wider population, these projects will be technology. The European Environmental Agency able to fight the problem of air pollution in the UK. (EEA) reported that battery powered electric As ultimately, it is an understanding of the sources vehicles have a net positive impact on air pollution of pollution, causes and fluctuations that will in comparison to its petrol or diesel equivalent, enable cities to control air pollution and effectively and many cities are already leading an example curb climate change; findings that IoT can enable. by pushing to increase the usage of electric mobility. This, when combined with government www.comms365.com i40today.com 39

Manufacturers are facing an unenviable sea of challenges By Nick Pike, VP UK&I, Outsystems Manufacturers are facing an unenviable sea of challenges as they bid to realise the opportunities of Industry 4.0. The businesses achieving the greatest gains in productivity, growth and inventory control are those that are successfully leveraging data-driven insights, adopting digital platforms and developing enterprise and customer-facing apps that draw on that data to give them a competitive edge. However, a shortage of digital engineering skills, lengthy IT backlogs, and legacy technology act like an anchor dragging on the ship of progress. Navigating these choppy waters requires These four factors create an environment that is an evolution in mindset and the adoption of overstretched, risk-averse, and slow to adopt new tools that facilitate innovation without needing ways of working. Recent research we conducted uncomfortable levels of investment risk. found that 64% of manufacturing businesses complain about IT backlogs, and only 31% of Barriers to Digital Innovation these respondents said that the backlog had improved in the past year. The challenge of transformation is particularly acute for mid-sized manufacturers. Giant As a result, important innovation efforts, digital multinationals have vast resources to devote to transformation programs, and Industry 4.0 innovation, and small youthful businesses haven’t initiatives risk getting snarled-up in lengthening IT yet acquired significant legacy technological queues. Such delays pose the risk that more nimble complexity. Firms in the middle have the worst of competitors will overtake you. Moreover, slow both worlds and typically face four major barriers delivery can mean that new digital propositions are that combine to slow digital innovation: obsolete by the time they reach market. 1. Complex, slow-to-change ERP systems, Fail fast, fail forward especially those that have been overly customised. What was originally implemented What’s needed is a shift in mindset away from the years ago as the steady, reliable, integrated tried and tested continuous improvement practices suite, now takes too long to adapt, and the of the past. Digital innovation sits on the other end resulting legacy gridlock slows innovation. of the certainty spectrum compared to Six Sigma DMAIC (Define, Measure, Analyze, Improve, and 2. The scarcity of web and mobile development Control). After all, there’s not much to measure and skills: according to the OutSystems State analyze if you’re really innovating, as opposed to of Application Development report, 82% improving a process that already exists. of manufacturing businesses suffer from a scarcity of application development talent. Manufacturers, many with decades or even Hiring and retaining developers takes longer centuries of history, are now being challenged and costs more. to act like start-ups and adopt an agile approach that follows the “fail fast, fail forward” maxim. For 3. A collaboration shortfall between IT, an industry that’s grown accustomed to lengthy engineering, and business staff: traditional and costly ERP implementations, where even IT development tools and practices come up minor changes seem to take forever, and cost a short when the business needs fast, iterative, fortune, the idea of experimental IT that’s fun, fast co-creative experiments. This, combined and low-risk must seem fanciful. with long backlogs, can lead to business and engineering staff taking matters into their One of the solutions to this conundrum is low- own hands, giving rise to poorly controlled code application development. Low-code shadow IT developments. While these might rapid application development platforms allow fill short term gaps, they can easily lead to manufacturing businesses to build robust support headaches, high costs, and security enterprise-ready mobile applications up to ten times vulnerabilities, especially if IT is kept in the dark. faster than traditional coding and makes it possible for a much wider range of employees to collaborate 4. Risk-averse IT sourcing practices: when with IT in jointly-staffed development projects. IT management has a lengthy backlog, it’s inevitable that they safeguard their precious With respect to the four challenges of digital resources. In that climate, getting buy-in innovation mentioned before, the contrast of low- for risky experiments is hard. That appears code rapid application development is that: to be borne out in the OutSystems State of Application Development report, which found • Typical projects take 6 – 12 weeks, instead of that manufacturers’ agile maturity lagged multi-month or multi-year ERP projects. behind that of most other industry sectors. 40 An insight into Industry 4.0 Sep/Oct 2019

• Engineers, business analysts, and process existing systems, with less reliance on lengthy ERP specialists can be quickly trained to help customizations. The result is quicker, and cheaper with application development, instead of to change as requirements evolve, which gives competing for increasingly expensive hard-to- manufacturers more control of their destiny. hire coding talent. A case in point was that of an imaging hardware • The collaboration gap between business and manufacturer that had scoped an 18-month project IT is bridged by forming co-creative teams to customise and integrate CRM capabilities into that blend IT, engineering, and business folk, its manufacturing execution system. Fortunately, who all use a fast, visual, and understandable before starting the project, the company did approach to application development. This a four-day proof-of-concept with OutSystems. reduces the risk of shadow IT. During those four days, using just one resource, they built 50% of the functionality required, and it • The firm’s appetite for digital experiments opened their eyes to the fact that these lengthy increases. When development is up to 10 times customisation and migration projects can become faster, and no longer solely dependent on scarce a thing of the past. developers stuck behind lengthy backlogs, the risk/reward dynamic completely changes. Looking to the future – bringing AI to bear Fast ROI makes risk-averse leadership teams more Adding AI into the mix can help manufacturers willing to embrace a culture of experimentation. to innovate and build new apps even faster. Often, we see low-code used in conjunction with OutSystems new application development, AI an innovation lab-- a safe place where engineers co-pilot, suggests next-best-steps to developers and IT can experiment with new technologies, based on a deep learning analysis of over 12 where the appetite for risk is so much higher, million anonymised development patterns. thanks to the high speed and low-cost paradigm of low-code software development. Results from the early-access-program include 25% faster development for experienced Regaining control of software development developers, and new OutSystems developers become proficient more quickly, thanks to the A frequent criticism of manufacturing software interactive AI assistance. systems of the past is the lock-in and cost associated with large ERP systems that need Developments like this are yet another way that extensive configuration and consulting, to adapt manufacturers can de-risk technology investment to the specific needs of an individual company. and counter the digital innovation deficit to become Low-code development, on the other hand, means faster, more agile and more open to new approaches that instead of buying and endlessly customising that will deliver the competitive edge essential to an industry package, it becomes feasible for survival in today’s commercial environment. businesses to either custom build, or extend, www.outsystems.com i40today.com 41

Consumers lack trust in artificial intelligence By i4.0 Today Editor Consumers lack trust in artificial intelligence (AI) and don’t understand the extent to which it can make their interactions with businesses better and more efficient, according to new research from Pegasystems Inc., the software company empowering digital transformation at the world’s leading enterprises. The study, which was conducted by research firm Savanta and unveiled at PegaWorld in Las Vegas, surveyed 5,000 consumers around the world on their views around AI, morality, ethical behavior, and empathy. Only 25 percent of Despite AI delivering the types of customized, • Most believe that AI does not utilize morality or consumers would trust relevant experiences people demand, many empathy: Only 12 percent of consumers agreed a decision made by an consumers still aren’t sold on the benefits. that AI can tell the difference between good and With many businesses turning to AI to improve evil, while over half (56 percent) of customers AI system over that the customer experience, it’s important for don’t believe it is possible to develop machines of a person regarding organizations to understand their customers’ that behave morally. Just 12 percent believe their qualification for perceptions, concerns, and preferences. Key they have ever interacted with a machine that findings of the study included: has shown empathy. a bank loan • Consumers are cynical about the companies One of the critical ways organizations can increase they do business with: Sixty-eight percent of customer trust and satisfaction is to use all the respondents said that organizations have an tools at their disposal and demonstrate more obligation to do what is morally right for the empathy in their interactions. But empathy is not a customer, beyond what is legally required. common corporate trait – especially when trying to Despite this, 65 percent of respondents don’t maximize profitability. As AI becomes increasingly trust that companies have their best interests important in driving customer engagement, at heart, raising significant questions about companies need to think about how to combine how much trust they have in the technology AI-based insights with human supplied ethical businesses use to interact with them. In a world considerations. that purports to be customer centric, consumers do not believe businesses actually care about To help improve empathy in AI systems, Pega is them or show enough empathy for their today announcing the launch of its Customer individual situations. Empathy Advisor. For further details on how this feature provides businesses with an ethical • There are serious trust issues with AI: Less framework to operationalize empathy and ethics in than half (40 percent) of respondents agreed all customer interactions please visit: https://www. that AI has the potential to improve the pega.com/ai-and-empathy customer service of businesses they interact with, while less than one third (30 percent) felt “Our study found that only 25 percent of consumers comfortable with businesses using AI to interact would trust a decision made by an AI system over with them. Just nine percent said they were that of a person regarding their qualification for ‘very comfortable’ with the idea. At the same a bank loan,” said Dr. Rob Walker, vice president, time, one third of all respondents said they were decisioning and analytics at Pega. “Consumers concerned about machines taking their jobs, likely prefer speaking to people because they with more than one quarter (27 percent) also have a greater degree of trust in them and believe citing the ‘rise of the robots and enslavement of it’s possible to influence the decision, when that’s humanity’ as a concern. far from the case. What’s needed is the ability for AI systems to help companies make ethical • Many believe that AI is unable to make decisions. To use the same example, in addition unbiased decisions: Over half (53 percent) of to a bank following regulatory processes before respondents said it’s possible for AI to show making an offer of a loan to an individual, it should bias in the way it makes decisions. Fifty-three also be able to determine whether or not it’s the percent also felt that AI will always make right thing to do ethically.“ decisions based on the biases of the person who created its initial instructions, regardless of “An important part of the evolution of artificial how much time has passed. intelligence will be the addition of guidelines that put ethical considerations on top of machine • People still prefer the human touch: Seventy learning. This will allow decisions to be made percent of respondents still prefer to speak to by AI systems within the context of customer a human than an AI system or a chatbot when engagement that would be seen as empathetic if dealing with customer service and 69 percent made by a person. AI shouldn’t be the sole arbiter of respondents agree they would be more of empathy in any organization and it’s not going inclined to tell the truth to a human than to an to help customers to trust organizations overnight. AI system. And when it comes to making life and However, by building a culture of empathy within a death decisions, an overwhelming 86 percent business, AI can be used as a powerful tool to help of people said they trust humans more than AI. differentiate companies from their competition.” Pega surveyed 5,000 consumers on their views on artificial intelligence, morality, ethical behavior, and empathy. The results included responses from the United States, the United Kingdom, France, Germany, and Japan. 42 An insight into Industry 4.0 Sep/Oct 2019

AI With Heart: Pega Launches 2. Predicting the Return on Empathy (ROE) First Customer Empathy Controls for Business Of course, no business can survive without turning a profit. Before putting empathy into action, Pega Customer Empathy Advisor helps can predict the effect more empathetic strategies businesses tune AI to forge more will have on their bottom line in real currency mutually-beneficial relationships terms. Businesses can then simulate scenarios at different empathy levels to see how that number Pegasystems Inc. today at PegaWorld introduced changes. From a profitability perspective, the Customer Empathy Advisor – a new capability system analyzes factors such as: of Pega® Customer Decision Hub that helps companies use AI to build more sustainable • Value to company: Is the action, offer, or customer relationships. The feature allows suggestion likely to increase the customer’s organizations to increase the level of empathy in lifetime value? AI-assisted conversations so they can build more trust, loyalty, and value with each customer. • Risk to company: Will it likely cause customer churn or increase exposure to risk? AI works quietly behind the scenes in all types of customer interactions – from marketing to sales • Compliance: Is it deliberately misselling a to service. Yet despite businesses’ best intentions product or unable to explain it? to use AI to make customers happier, most AI just adds more spam to people’s lives, sacrificing long- 3. Putting empathy to work term loyalty for a few short-term sales. Now as AI becomes easier to deploy, the debate on the With the right balance of empathy in place, ethics behind it has only intensified. For example, businesses can automatically deploy these just because AI could likely sell a high-interest loan strategies through Pega Infinity™ and its suite of to a low-income family doesn’t mean it should. customer engagement software – all powered How can companies use AI to balance the best by the same centralized AI from Pega Customer interests of both the company and the customer? Decision Hub. This means empathy can be applied in all marketing, sales, and customer For the first time, organizations can now service interactions, both online and off. Instead operationalize empathy at scale for the mutual of force feeding a particular strategy to a benefit of all involved. Customer Empathy Advisor segmented audience, companies can optimize provides businesses with an ethical framework to the empathy levels in the strategy best suited to instill empathy in all customer engagements and the needs of each customer – which in some case measure the effect. At its core, the AI-powered may mean taking no action at all. Pega Customer Decision Hub analyzes customer data to guide customer-facing agents and virtual The Customer Empathy Advisor will be available assistants to take the next best action with each for all Pega Customer Decision Hub clients near customer. Its Customer Empathy Advisor feature the end of 2019. It will also be demoed and on takes this analysis one step deeper by examining display this week at PegaWorld, Pega’s annual how empathic these recommendations are and customer conference now being held for more offers a more compassionate approach – which, than 5,000 digital transformation professionals at contrary to popular belief, is often the most the MGM Grand in Las Vegas. profitable strategy as well. Pega Customer Decision Hub enables Pega Customer Empathy Advisor takes a three- organizations to surface unique insights and step approach to integrating empathy into recommend the next best action in real time on enterprise decision-making: every step of the customer journey. It provides the centralized AI power across the unified Pega 1. Analyzing empathy in any engagement Infinity™ digital transformation software suite, which optimizes customer engagement and First, Pega’s machine learning capabilities analyze operational efficiency from end to end on a global the organization’s marketing, sales, and service scale. strategies to assess the current level of empathy. The Customer Empathy Advisor dashboard “Most businesses are conditioned to try and breaks down the different elements of empathy squeeze every last drop of profit from each applied within each strategy, including: customer. But this predatory mentality distorts the fact that practicing a little empathy is not only • Relevance: Is the action, offer, or suggestion good for the customer, it’s good for business,” said of interest to the customer? Dr. Rob Walker, vice president, decisioning and analytics, Pegasystems. “We’ve always believed • Suitability: Will it likely cause harm to that the only way to win a customer’s heart is to the customer? first walk a mile in their shoes. Today we’re taking a step closer in this pursuit by instilling empathy • Value: Will the customer likely benefit from it? in customer interactions – which is ultimately the right way to do business for everyone.” • Context: Is it consistent with the customer’s recent activity? Supporting Resources • Intent: Does it take into account the • Feature background: Customer Empathy Advisor customer’s likely goals? • Analyst report: Forrester Names Pega Leader in • Mood: Does it align with the customer’s Real-Time Interaction Management current frame of mind? • Research: Consumer survey on attitudes on From there, engagement analysts can pinpoint empathy in AI which actions engender the most trust from customers and which repel them farther away www.pega.com from the brand. Using a simple intuitive slider, stakeholders can easily adjust empathy levels and see the impact the change has on relevant KPIs. i40today.com 43

OMRON contributes to resolving the world’s social issues through its businesses By Kevin Youngs, European Sales Manager Automated Inspection Systems Division, Omron Europe B.V. Omron is listed in the Forbes top 2000 largest companies in the globe, with more than 80 years’ experience in sensing and control. We’ve come a long way since Kazuma Tateisi founded our company in Osaki in 1933. More than 12,500 issued and pending patents underline our innovation strengths. By weighing the economic, environmental and Since OMRON joined the PCB inspection system social impacts, we underline that people are business in 1987, it has maintained the No. 1 important to us with over 37,500 employees position in the industry due to advanced inspection worldwide, spread out over our 200 locations, performance and strong robustness, providing Omron annual sales are quoted at 6.1(Billion Euro), high yields and lower manufacturing costs. which is divided under Industrial Automation Solutions (including Automated Inspection Systems Omron annual sales Division), Components, Automotive Components, are quoted at 6.1(Billion Social Systems and Healthcare Solutions. All Euro), which is divided contributing to a safe society in which we live. We provide everything for manufacturing solutions under Industrial Input, Logic, Output, Safety, Robot Solutions and Automation Solutions Automated Inspection Systems. (including Automated Inspection Systems In addition to manufacturing solutions, Omron Division), Components, have over 12 manufacturing facilities world- Automotive Components, wide; three main Electronic Manufacturing sites Social Systems and in Japan (Kusatsu and Ayabe), China (Shanghai) Healthcare Solutions. and in the Netherlands (Den Bosch). Combined with our wealth of innovation technologies, our knowledge of Electronic Manufacturing processes enables us to develop products and technologies from the forefront of experience. Our approach to resolve social issues are one representation of how OMRON is a corporate group that people can always depend on, and an organization that continues to live up to the high expectations of people all over the world. For visible (optical) inspection, Omron has achieved hybrid 3D measurement technology, which combines Color-Highlight and phase shift technologies called 3D-SJI. By combining the captured solder shape with high specularity through Color-Highlight and accurate height measurement through phase shift measurement, 3D-SJI accommodates variations of the gloss and shape of the solder surface. For non-visible (X-ray) inspection, the high- accuracy 3D computed tomography (3D-CT) technology is adopted. One challenge was the achievement of high-speed inspection practical for inline inspection. To overcome this challenge, 44 An insight into Industry 4.0 Sep/Oct 2019

Connecting Global Competence November 12–15, 2019 Accelerating Innovation co-located event World’s Leading Trade Fair for Electronics Development and Production November 12–15, 2019, Messe München productronica.com i40today.com 45

Omron has developed an inspection principle advanced technological “knowhow”, OMRON has called parallel CT imaging. In this principle, the cultivated in automating the solder inspection. X-ray camera is moved horizontally without rotation in the XY stage imaging system to OMRON has been intensifying our efforts in secure a large visual field size that can be used the SMT process to automate the dimension & for inspection, making it possible to reduce the visual inspections of PCB’s. With our customers, number of FOV’s (field of views), which becomes OMRON will continue to aim for “ZERO DEFECT a factor to increase inspection time. – constructing production lines that eliminate defective products” and provide inspection Combined with our systems that can contribute towards achieving wealth of innovation this with advanced innovations in Post Reflow Inspection with 3D-AOI and 3D-AXI utilizing high technologies, our speed Computed Tomography particularly in knowledge of Electronic Automotive manufacturing. Manufacturing Along with changes in the trends of the world, processes enables us to “manufacturing” has also undergone a major develop products and change with significantly increased quality level technologies from the requirement. To support 4Ms “Man worker)”, forefront of experience. “Machine (facility)”, “Material (raw materials)”, and “Method (work method)” and to continue OMRON Inspection Systems continuously manufacturing high-quality products, Omron endeavour to improve customer quality and proposes Q-up System ‘Smart Factory’ solution productivity, with the quality information of the utilizing AI that performs linking of data from inspection system as a starting point. Back up with facilities, collection and monitoring; also automatic feedback and feedforward of the manufacturing data and inspection result data with the goal to improve manufacturing quality and yields, based on the quality information of the inspection system. This allows real-time capturing of “changes that affect quality” of 4Ms in production fluctuations, and assisting efficient improvement activities particularly in Automotive manufacturing. www.omron.com 46 An insight into Industry 4.0 Sep/Oct 2019

register at advancedelectronicsassembly.com In association with and Budapest, Hungary 19th November Oradea, Romania 21st November A VIEW OF THE LATEST INDUSTRY TECHNOLOGIES improve yields • reduce costs • overcome challenges advancedelectronicsassembly.com 47 i40today.com

Current partners Koh Young America, Inc. Find us at 1950 Evergreen Blvd., Suite 200, Duluth, GA 30096 +1-470-374-9254 | [email protected] BOOTH #609 kohyoung.com