WashU Engineering-designed autonomous car inspired by insects

After about nine months of tinkering in the lab — and only a month's notice — a team of five School of Engineering & Applied Science students placed among the top 10 teams and won $750 in an international robotics competition in China for their low-cost autonomous vehicle design inspired by the movements of insects.

A team of five students placed among the top 10 teams in an international robotics competition for their low-cost autonomous vehicle design inspired by the movements of insects.

Adith Jagadish, a master's student in robotics; Matt Kollada, who earned bachelor's degrees in systems science & engineering and finance in 2017; Will Luer, who earned a bachelor's in mechanical engineering in May 2017 and is now earning a master's in computer science; Andrew O'Sullivan, a senior majoring in mechanical engineering; and Meizhi Wang, a BS/MS student in electrical engineering; comprised the team that traveled to Xi'an Jiaotong University in Xi'an, China, the first week of June for the Silk Road Robotics Innovations Competition, in which 34 teams competed.

The WashU Engineering team, the only team from the U.S. in the competition, learned of the event in early May, giving them about a month to prepare.

Members of the team were working to develop a low-cost, self-driving vehicle design in the labs of Xuan "Silvia" Zhang, assistant professor of electrical engineering, who accompanied the team to China as faculty adviser.

"One of the original objectives of this project was to see if we could make this in a very affordable manner," Zhang said. "You can buy an expensive robotic experiment setup that's ready-made, but to make this affordable, on one hand, it presented some interesting constraints that we had to work with as engineers. With extra constraints, there is more room for innovation and design. On the other hand, we want to be able to open source the entire platform so that other people can replicate it. In the end, that's a limitation, but we also tried to embrace this as a challenge."

The team's autonomous car, called the FlowBot, is about 4 inches by 7 inches, is inspired by the way insects use an optical flow method to navigate and avoid obstacles in an agile manner. The base model is called the Pi car because it uses Raspberry Pi, a palm-sized computer that costs about $30; a Raspberry Pi camera that costs about $15; and two motors.

The team began building the Pi car as a model on which others could use to build their own autonomous car using the team's code, which is available on GitHub. But the car took some time to get running.

"The lab started working on this robot last fall, but we never got it working — the car never ran, and it never moved," O'Sullivan said. "We worked on the code and did some 3-D printing and talked about the best way to drive the car. As soon as the competition was in our minds, we were working 40 hours a week on it. We did over a semester's worth of work a week for several weeks, which is what allowed us to come together to prepare for the competition."

At the competition, which included seven teams from countries outside of China, each team had to give a 1-minute video presentation followed by a 5-minute PowerPoint presentation. Teams that made it to the next round had to give another presentation then demonstrate the robot on the stage for the judges.

Ultimately, the team made five cars, each slightly different to show different features. They each carried components in their luggage to China. When they arrived, they had to rebuild and debug the cars in borrowed labs at Xi'an Jiaotong University, a McDonnell International Scholars Academy partner. The team worked up until the minute of their presentation, O'Sullivan said.

"Our robot gave a great showing," O'Sullivan said. "We set up a minefield of water bottles on the stage. It was able to dodge all of the bottles four times in a row, except it hit one bottle, which counted against us."

"This competition, and many like it, demonstrates that making artificial intelligence technology practical and useful for robotics and various Internet of Things applications is not about building a car that is 100 percent reliable or a robot that never bumps into things, but rather about how to develop a system that is resilient and self-healing," Zhang said.

The McKelvey School of Engineering at Washington University in St. Louis promotes independent inquiry and education with an emphasis on scientific excellence, innovation and collaboration without boundaries. McKelvey Engineering has top-ranked research and graduate programs across departments, particularly in biomedical engineering, environmental engineering and computing, and has one of the most selective undergraduate programs in the country. With 140 full-time faculty, 1,387 undergraduate students, 1,448 graduate students and 21,000 living alumni, we are working to solve some of society’s greatest challenges; to prepare students to become leaders and innovate throughout their careers; and to be a catalyst of economic development for the St. Louis region and beyond.