What will drive test & measurement?

-September 08, 2016

At the start of EDN's 60th anniversary summer, I posed several questions to people around the test industry, asking for their thoughts on what will drive test in the coming years. Two people gave such detailed answers that I've decided to combine them and share it with you. What's interesting is how these two, from different generations, responded to the questions.

Bob Witte is Vice President of Technology Strategy for Keysight Technologies. Bob has worked in a variety of test and measurement product lines, mostly in R&D assignments. He is the author of Electronic Test Instruments: Analog and Digital Measurements and Spectrum and Network Measurements, and is a Senior Member of IEEE. He holds a BSEE from Purdue University, an MSEE from Colorado State University, and an MS in Management of Technology from the National Technological University.

Shayan Ushani is the co-founder Orenda Financial, and a staff writer and marketing & business director of the Archway newspaper. He is a student at Bryant University with triple concentrations in finance, economics, and applied analytic. Shayan helps with the business development of Analog Arts, managing social media and editorial aspects. His interests lie around FinTech, Big Data, and the data revolution. He has also written several articles for EDN including Bandwidth 101, FFT: Equations and history, and The Art of Waves: Images on an oscilloscope.

Will engineers need more from software? Is hardware a commodity and all the difference is software?

Witte: Clearly, the balance between hardware and software contribution is shifting as the software content increases in T&M systems. We might see hardware become a commodity in the future, especially for less demanding applications, but there are still significant measurement challenges that can only be achieved with high-performance, usually custom hardware. For example, a 100 GHz oscilloscope is a state-of-the-art design challenge requiring custom ICs designed using the fastest semiconductor processes available.

Ushani: Software is the key to the progress of any high-tech device and, in general, the future technological revolutions. Clearly, the test and measurement industry is no exception. Imagine a conscious system that knows how to configure itself to test anything similar to a neural network employing machine learning, while the operator is far away. Newer programming languages with more data processing efficiency will be introduced and utilized in the industry. We can see that already happening with MATLAB.

Will IoT change anything or is it all just an integration of existing technologies? Will most of the IoT testing be for testing software, as in for security?

Witte: The IoT trend is a great example of "all the technology exists" but the pervasive adoption of it will produce dramatic shifts for many industries. For the T&M industry, it is likely to drive change in several fundamental ways:

  • Our T&M products will be IoT devices, collectors of data that fit into big data systems. We already see this now with Ethernet-enabled instruments and it will continue to expand.
  • The cyber security issues associated with IoT are huge and currently unsolved. Dramatic improvement in this area is required. Access to IoT data must be controlled but cannot be too cumbersome, which will impede broad deployment.
  • Many of the expected IoT devices will be low cost which will put pressure on the cost of test in manufacturing. In some cases, BIST may be sufficient to meet customer quality requirements. In other cases, functional test will be used and it must be fast and cheap.

Ushani: IoT and predictive analytics will definitely play a big role in the future of test and measurement. Data loggers will exchange information to prevent disasters like the Fukushima Accident. Robots will exchange T&M data to stop a heart attack or fine tune a musical instrument. From wearable technology to remote tracking systems; artificial neural networks will be implemented in T&M networks.

Everyone is talking 5G now, but what might be next?

Witte: 5G is a big challenge that is consuming the time and attention of all the players in the wireless market. It will take time to fully roll out worldwide and will not be delivered in one big release, with major network deployments schedule for 2020 and beyond. After that, who knows? But, the demand for more bandwidth and high-availability connections will continue. There has been some speculation that 6G wireless will integrate satellite-based systems into the network to address the global coverage problem.

Ushani: The rapid progress from 3G to 4G to 5G, was mostly due to the speed factor, but what will revolutionize the test and measurement world will be how the information is transferred. As we have been seeing, encryption is an important key in data processing. We also believe that more efficient data compression technology will further advance the technology.

Does test cost too much?

Witte: In an ideal world, test does not add any value and should be eliminated. If we had perfect design, perfect parts, and perfect manufacturing processes, products could just roll off the production line without any test. Of course, we don’t live in that perfect world, so test remains an essential part of the manufacturing process. Still, electronics manufacturers strive to reduce test costs to remain competitive. Robust manufacturing processes incorporate measured data to optimize yield, cost, and performance (e.g., Industry/Industrie 4.0).

Ushani: Automation could result in less overall cost for testing as test instruments become more efficient and powerful. As test equipment becomes more capable and sophisticated, it may cost more.

Will the traditional test engineer be a job of the past?

Witte: Software development is clearly becoming a bigger part of the test engineer's job. Like many engineering jobs, the skills required for test engineering are trending toward a combination of interdisciplinary skills: metrology, measurement science, software development, and systems engineering. For many industries, time-to-market pressure and the need to reduce test-development costs mean that test engineers need to create test systems quickly and at low cost. This implies an increased use of off-the-shelf software components and measurement libraries. The evolution of manufacturing processes (Industry 4.0) will drive the test engineer's work to be tightly integrated into the overall manufacturing flow (not just a pass/fail test at the end of the line). Machine learning is emerging as a powerful technology for future manufacturing systems and test engineers will need to be building new skills in that area. So the job will change but likely not disappear.

Will engineers need more software skills than they have today? Less emphasis on hardware skills? More interdisciplinary skills?

Witte: Some level of software skill is required by all engineers. Even if they don't "write code" as a main part of their job, knowing how to create and manipulate software is an extremely valuable skill. Engineering schools are generally doing a good job of making sure that software is present in their core curriculum. We will continue to see the software curriculum mature, reflecting that there are multiple disciplines included under "software" such as human interface design, application layer, architecture, hardware control, signal processing, etc.

Ushani: Software developers will become a norm even in the hardware arena. In the future, engineers need to constantly refresh their skills as the industry moves along at a fast pace. They will also have access to more resources and tools, which could demand some form of education and learning process. For example, various software translators could become a norm in the near future.

Which industries might drive test: wireless, consumer, telecom/datacom, medical, automotive, military, aerospace?

Witte: In general, the markets with the highest growth and investment tend to drive opportunity and technological innovation in T&M. Clearly, the 5G investment fits this model with a huge investment and some very difficult test challenges in all lifecycle phases: early R&D, design and development, deployment and maintenance.

Investment in wireline networks continues strong due to increasing bandwidth requirements driven by customer demand. Automotive is another strong area as the electronic content in vehicles continues to grow. Aerospace/defense applications continue to see investment, but with lower overall growth rates. These applications, however, tend to require higher frequency and higher performance, which pushes the performance of test-and-measurement gear.

Ushani: A lot of technology has sprouted from our defense industry. The U.S. Navy built the first "prototype" of what is now today's World Wide Web and the Air Force led to the successful development of our trek into space. As before, the military will be in the forefront of industrial revolution.

Will regulations force more testing?

Witte: Regulations drive a certain amount of test activity (EMC, for example) but industry standards are a bigger factor. In a highly connected world, emerging interface standards must be "tested to spec" so that products from different vendors interoperate properly. This is true for a wide variety of standards, including USB, HDMI, 100G Ethernet, and 5G wireless. We work closely with the standards bodies to ensure that the specs are written to be testable and that validation tests are in place when a new standard is deployed.

Ushani: As robots become popular, safety regulation will become an important issue. A federal agency, similar to FDA, could pop up to oversee all that is coming.

What technologies will emerge from today's research?

Witte: Machine learning, big data, and cloud computing are emerging technologies that have the potential to drive major changes in how electronic products are developed and manufactured.

Ushani: A big technology that could advance from today's research is biometrics, the use of biotechnology and synchronization with the self. T&M would be implemented to constantly test and monitor a person with a certain condition, like diabetes. The IoT and applications of big data will allow it to self-administer to the patient. From a patient who needs insulin when his blood sugar is abnormal to others who have just missed a routine checkup. Other industries, which could benefit from today's advances are machine learning and entertainment.

More predictions:

Also see:

Loading comments...

Write a Comment

To comment please Log In