Homing in on STEM Innovation

High-school senior Victor Cai’s short-range, narrow-bandwidth radar project netted him a $175,000 science award – and could be useful for future AVs.

High-school senior Victor Cai with his innovative radar set; the tin cans function as antennae. (Society for Science)

Constraints are sometimes just new ways to open doors. High school senior Victor Cai (above), an 18-year-old from Lehigh Valley in Pennsylvania, learned this lesson working on a short-range, narrow-bandwidth radar project that ended up taking second place in the 2022 Regeneron Science Talent Search run by the Society for Science . This year, 1,804 students from 603 high schools in 46 states entered the contest.

Cai was awarded $175,000 for his radar design, which transmits two signals at different frequencies and then measures the phase difference between them to calculate distance. Cai used a multi-frequency continuous wave (MFCW) radar algorithm and software instead of specialized hardware in his design. Cai’s radar system is accurate to within 12 centimeters (4.7 in.) while also needing five magnitudes less bandwidth than standard radar. The Society for Science said technology like this could be useful in self-driving vehicles that can do more with less.

Cai’s winning radar is the latest in a string of science projects he created in middle school and high school, including a hydraulic power miniature crane built with syringes. Cai said seeing a Doppler radar project at another science fair got him thinking about how he could create a distance sensing radar using similar concepts. “When you think of radar, you usually think of long range, distance sensing radar,” he said. “I wanted to make a short-range radar that I could test within my own home. It turns out, that was a lot harder than I had expected.”

Narrow bandwidth

Cai only had four megahertz of bandwidth to work with for his project because of the software defined radio (SDR) technology component he had access to. “Using a traditional radar algorithm, four megahertz of bandwidth would give you a theoretical resolution of 37 meters,” he said. "In other words, don't even try to look for something in the short range.”

Undeterred, Cai looked around for alternative algorithms, and eventually found a MFCW radar algorithm that offered the possibility of making a short-range, distance-sensing radar that could operate on his SDR. “I was able to get that down to only a 10-kilohertz bandwidth usage, which is well within the SDR’s four megahertz of bandwidth,” he said, adding that he knew he was also be constrained to the short WiFi band (from 2.4 to 2.5 gigahertz) indoors, which is only 100 megahertz wide. “Working within those limitations, I was able to find a way to make short range distance sensing work,” he said.

Cai’s MFCW algorithm uses two frequencies as pure sine waves (one at 2.4 GHz, the other a 2.5 GHz). It uses the phase difference between the two frequencies, which grows linearly with the actual distance, to detect how far away something is. “And then for a Doppler, you can simply take one of those two channels and use that continuous wave,” he noted. “And since there’s a very slight frequency shift, when it comes back, that will give you the velocity information.”

The limits Cai faced in his student project are exactly the reasons why this work could have implications beyond a science fair. Increasing the bandwidth would have required him to either buy or make custom hardware or he would need a more-expensive SDR. The idea of using lower-cost, lower-bandwidth sensors in mass-market autonomous vehicles certainly makes sense, if Cai’s tech can someday find its way into production AVs. The multi-talented student (he is also an accomplished viola musician) himself has not yet ridden in an autonomous vehicle. He said he would likely feel “slightly scared” on his first time.

Needed applications

Commercial application of Cai’s project, which uses tin cans as antennae, is not in the immediate future. But there’s room for creative thinking like this in the auto industry, said Peter Seydel, senior expert for radar technology at Continental. Seydel didn’t comment on Cai’s project specifically but did say AV adoption can only be helped by these kinds of projects.

“Evaluating novel approaches to radar design is important to broaden [AV] market penetration,” he said. "Getting more advanced technology into vehicles on a broader scale needs continuous improvement in affordability, vehicle integration flexibility, and so much more. Continental encourages students to explore innovation within the automotive space.”

Cai wasn’t thinking about autonomous vehicles when he was working on his radar – the original plan was to create a sort of “digital guide dog” for his visually impaired karate teacher. But now that he’s got the concept down, he did look into possible AV applications, and realized there’s still work to do.

“There's like a whole four gigahertz-wide bandwidth, somewhere up in the 80 gigahertz range, dedicated for autonomous vehicles and vehicle radars,” he said. “Those vehicles are using a whole gigahertz of bandwidth each time they're using it. And I think it's very important, especially if once autonomous vehicles become popularized, that we're able to reduce that bandwidth requirement.” He reckons that will significantly reduce radar interference in dense traffic situations.

Cai said his prize money is going towards his college education. He’s considering schools like Princeton University, Yale University and the Massachusetts Institute of Technology for undergrad work and then finding a graduate school for electrical engineering. Cai said he’s interested in studying signal processing or wireless communication technologies, so don’t be surprised if some future AV ends up being a bit better thanks to a bright kid with some tin cans and an idea.