Skip to main content

Interview: Behraad Bahreini

Tell us about yourself and how you got to Vancouver?

Ever since I was a graduate student at the University of Manitoba, I have been working on microsystems. These miniaturized devices operate with small amounts of energy, and one needs suitable test equipment to study them. Measurement quality has always been a concern. I used lock-in amplifiers in many of my tests, not only for the improved noise performance but also to study nonlinear phenomena. I was in charge of developing interface circuits at Cambridge University. Later I worked for a company as an engineer before moving to Vancouver to take an academic role. For our research testing is part of the design process, so when we think about a new device, we include measurement considerations from the start.

You design devices for many application areas (on the website: IoT, Automotive, defense, biotech). What is your main focus?

Most of our works include resonators. However over the years the research focus has changed a lot. Maybe ten years ago we were looking at loss mechanisms at small scales and how they can be avoided. As the field progressed, we investigated the non-linearity and coupling of these resonators which is harder to understand. Our technological progress is motivated by fundamental studies, while the transfer to applications comes gradually over time. So it happens that communications, defense and automotive are the main focuses of our team.

What will the research trends be in the next 3-5 years in your opinion?

If for instance, you look at accelerometers in terms of absolute performance, there has been not much improvement in recent years. The limits of performance-to-cost on the physical side are reached in many cases. What is going to happen is increased investment in software and signal processing. The math and physics behind the operation of many such sensors have been worked out. With the availability of inexpensive computation, the research may evolve towards higher level of integration of multiple sensors (e.g. arrays) in order to collect more measurements per time unit and improve the performance through subsequent processing of those results. If you use parallel sensors your signal becomes stronger, and mathematically the noise floor goes down. Additionally, you can get gradients that give additional information.

Where does intelligence play a role with your sensors?

We have been working a long time on increasing the capability of our devices by improving mechanical or electronic designs. During my sabbatical in 2016 and having worked with the R&D of a company, where I tried to apply new concepts like sensor fusion and statistical signal processing to our designs. For us, this knowledge was partly new. Other people talk about machine learning or artificial intelligence when using such algorithms. But we are not data scientists and need to be careful with such terminology. However, we have evolved our sensors, since we know the physics, we know how these devices work, and now we have added some intelligence to extract more information from them.

What would be an ideal measurement instrument for you?

We like to have access to raw data. In many situations, this is preferred, and then we are ready to study the data. This is especially needed in the early stages of research. In other situations, software to treat the raw data is equally appreciated, which simplifies the related analyses. Then there are features enabled by the combination of hardware and software that we find essential, like multiple harmonic measurements. Actually, we like high performance (in terms of noise, dynamics, bandwidth) instrumentation that at the same time provides the flexibility to go beyond the original application. An example of this is the control of the data acquisition rate: short experiments (seconds) are as important as long measurements (weeks) without the need to end up with huge files. Therefore, we like features that provide flexibility.

What does your lab offer to new grad students?

Students come here to learn in a multi-disciplinary team. Moreover one needs to consider that the group is big and the opportunities for collaborations that derive from that are multifold. I can promise to our grad students that they will learn a lot with access to world-class facilities. We have truly multi-disciplinary research. After a couple of years in our lab, one learns about physics, mechanics, electronics, and signal processing. On the other hand, Vancouver is occasionally overtaken by Zurich and Vienna in the ranking for quality of living, but let's hold down that we've constantly been in the top 5 for the past 20 years. You will love Canada!

How do you achieve your work-life-balance?

I have decided that skiing would be a good way to end my life prematurely, so I did not invest into that. Being a professor, it is not easy to keep track of private interests, although some may do it better than others. For me it was essential to be there for my family and kids when the job permitted that. There are no set working hours in research, but definitely the freedom of choice to concentrate on what you want is worth the extra time I put in my career.

Behraad Bahreini

Behraad Bahreyni leads the Intelligent Sensing Laboratory at the Simon Fraser University in Vancouver, Canada

Read more interviews
Contact Us