Designing IoT devices to benefit society

Banner image: (L to R) David Radke, Amelia Holcomb, Omid Abari and Farzan Dehbashi.

A device that detects if someone is drowning in a pool to help relieve a parent’s worry. A wireless gadget to more quickly identify fires, so everyone gets to safety. A system to track and assist people with Alzheimer’s to help calm a family’s fears when they’re not present. These are just a few examples of projects designed for social good by students in the University of Waterloo’s Faculty of Mathematics graduate course, IoT and Intelligent Connectivity.

The course started in 2018, is the brainchild of Waterloo’s David R. Cheriton School of Computer Science assistant professor, Omid Abari.

“At the end of the day, the devices should have a benefit to society,” Abari emphasized. “That was my goal, and I think whenever students can see the real application for what they’re doing, they are more motivated.”

Master of Mathematics in Computer Science student, Farzan Dehbashi, appreciates that students’ inventiveness is encouraged, which resulted in his group coming up with SwimTrack, a system to detect drowning in swimming pools.

“One amazing feature of the course is that students are encouraged to think outside of the box and attempt challenging projects,” said Dehbashi. “This was a factor that encouraged us to work on SwimTrack as it required a swimming pool with all of the safety features that, at first glance, seemed to be far-fetched.”

SwimTrack includes RFID tags attached to the hand of each swimmer in a pool, and then a RFID reader placed outside of the pool simultaneously reads all the tags and can detect if swimmer is drowning or swimming. The tool can also identify the type of swimming each person is performing.

The course made such an impression on David Radke, who worked with his group to come up with the project called “Detecting Fire using Wireless Signals,” that he has decided to switch from pursuing a master’s degree to the PhD track.

“I really enjoyed the projects and the research in the course,” Radke said. “The research I began in the course, and have continued to work on after the course, helped me make my decision to switch from the master’s track to the PhD track. My main research focus will be on exploring different methods and developing new AI models for temporal analysis to answer time-sensitive questions across different domains, including IoT systems.”

Detecting Fire using Wireless Signals circumvents some limitations of today’s indoor fire detectors (smoke alarms, heat detectors, etc.) such as latency and range. Their method can enable an entirely new fire detector device to solve the limitations of existing methods.

Abari limited last year’s class size to 25 students with all the spots taken within a day. For this academic year, he has further reduced the class size allowing for only 15 students.

“I limited the class size because it’s a very hands-on course, and the students work together, and I want them to get the most out of it,” Abari explained. “When the class gets too big, students don’t discuss as much, and their involvement goes down. I wanted to make sure the class is a size where I can spend time with every single student.”

Students who take the course have the rare opportunity to learn hardware and software at the same time. At the beginning of the course, students are taught fundamentals about how to design IoT networks and IoT devices. Students also engage in discussions about state-of-the-art research in the cloud computing domain and what’s happening in the world in this area.

The second half of the course is when the students participate in hands-on projects — building devices and designing algorithm software for different applications.

The students, who are from diverse backgrounds, work in project teams (ideally three), with each of student contributing specific skills as they also learn from each other.

“The diversity in research backgrounds of students who took the course helped new ideas flourish that may not have happened if everyone had worked within the same research domain,” said Radke. “The papers we read were equally diverse, allowing students with particular backgrounds to help explain the work to other students who didn’t quite have the background in that area. Overall, the course was a great introduction to how aspects from different areas within computer science and engineering can work together to invent new ideas.”

First-year Master of Mathematics in Computer Science student, Amelia Holcomb, who is currently taking the course, is hoping to focus her research on applications of computer science for clean energy, climate, transit, or the environment.

“I was interested in this class because I want to work on applications of computer science to solve real-world problems; the climate in particular,” Holcomb shared. “So far, I’m really enjoying the class. I especially like that every paper discussion includes ‘What are other applications of this research? What else could we use this for?’ It’s a class where wild and wacky ideas are encouraged.”