It seems as if every time we’re approaching the boundaries of technology’s influence on our culture, a new development, trend, or exciting innovation pushes those boundaries even further, launching even further exploration. A “Digital Transformation” has been unfolding all around us over the past few years — the rapid shift to on-line banking, e-commerce, medical records, etc. — that is having a profound influence on our day-to-day lives.
Networking and wireless connectivity are enabling similar productivity gains in the workflow for scientific and engineering applications. Specifically, in the test and measurement world, there is a clear transition to more networked or “connected” devices, making it easier and faster to get data and measurements from remote devices and systems to the PC or “Cloud” for further analysis, better data management, and improved overall productivity. This movement is creating a demand for technology that I am going to refer to as “Embedded Measurements”. Let’s explore it together.
Where Will “Embedded Measurement” Be Relevant?
Real-time, continuous measurements are an important component of the larger digital transformation appearing in factory automation, transportation, energy and other industries . The IoT (“internet of things”) trend is enabled by embedded measurements, which are at the confluence of measurement, embedded design, and software. In the consumer products world, the ability of your thermostat or smart refrigerator to communicate status is the result of embedded measurements and cloud-connected devices working inside the device. This same capability is being applied to industrial applications as well, from automobiles and security systems to machinery and medical devices.
Some of these applications might require much higher measurement speed and accuracy and considerably more processing power than what is afforded by the low-cost devices in consumer products. In these applications, real-time digital signal processing is often required to aggregate and process large amounts of data near the sensor or device (“at the Edge”) before sending updates to the cloud. This is where embedded measurement systems shine — they are the convergence of benchtop instruments with the capabilities and characteristics of general-purpose embedded computers — network connectivity, stand-alone processing, user programmability, etc.
Why use FPGAs in Embedded Measurement Systems?
FPGAs bring a unique combination of performance, parallelism, and power efficiency. This combination has made FPGAs the computational technology of choice for digital signal processing (DSP), largely replacing general purpose DSP devices. Since signal processing is fundamental to test and measurement applications, FPGAs are being used extensively in new instrumentation designs, from general purpose test and measurement products to more application-specific instruments in medical, semiconductor, aerospace, and other industries.
More recently, the natural fit of FPGAs to neural nets has helped accelerate the emergence of artificial intelligence and machine learning (AI/ML) from the lab into commercial applications. This rapidly growing trend is putting more sophisticated processing and decision making near the sensor or “at the edge”, improving responsiveness and reducing demands on network bandwidth and reliability.
Software also plays a critical role in Embedded Measurements, enabling engineers to easily acquire, save, manage, process, and visualize their measurement data and results. The choices customers have for automating their test and measurement systems are growing, with newer entrants like Python adding to established favorites like MATLAB and LabVIEW. Digilent bridges between the embedded systems world with lower-level languages focused on efficiency while providing the productivity advantages of higher-level languages to connect devices to each other and the cloud.
What Does This Mean for Engineers?
Since measurements are increasingly inherent in the operation of a device, especially as devices become “smarter” and more aware of their own status and the environment they are operating in, electrical engineers will have to be better at understanding how to interface with sensors and process data. This has historically been an area of specialization (instrumentation and data acquisition), but even electrical engineers that have historically focused on digital electronics will need a good understanding of how to incorporate measurements into digital systems with efficiency and accuracy.
A Final Word
Engineering is fluid. It’s always been straddling the line between ‘Progress’ and ‘Do What Works’ — it’s about putting new technology into action. As we look to 2021 and beyond, I believe that the most interesting progress will be made where traditional engineering disciplines overlap with each other: mechanical, biomedical, electrical, RF, computer science. Part of the challenge (and why I love being an engineer) is being comfortable trying something new and bridging those traditional boundaries. Embedded measurements are an incredibly enticing opportunity for myself and Digilent, and my hope is that through a line-up of exciting products we have planned for this year along with our first products in this area (Eclypse Z7 and USB104 A7), there will be plenty of tools to keep my fellow engineers engaged as they navigate this new frontier.