Self-driving cars are now a booming industry, but that boom has come with growing pains. For example, where does an aspiring autonomous car engineer go for job training? The University of Toronto hopes to fill that void by teaming up with online education company Coursera to offer a specialization in self-driving cars.
“It’s pretty clear that it’s an exploding area, and there’s not a lot of introductory material,” University of Toronto Professor Steven Waslander told Digital Trends about the rationale for starting this program.
The specialization is designed for students with an engineering background, but little to no formal training in self-driving cars. The University of Toronto will offer a series of online courses covering important topics, such as how to build the digital maps that guide autonomous cars, computer vision, and how to write the software that governs cars’ movement. By the end of the four-course specialization, students will be able to send a virtual car around a simulated track, according to Coursera.
Much of the development work on autonomous driving tech is being done by private companies, from automakers like Ford and Toyota to tech companies like Waymo and Lyft. These companies carefully guard their work in the name of achieving competitive advantages.
The University of Toronto can’t get behind those closed doors (although autonomous-driving companies Oxbotica and Zoox will provide some input), but it can provide a good overview of how self-driving cars work, Waslander said. He noted that the technology originated in academia with programs like the DARPA Grand Challenges. In his previous post at the University of Waterloo in Ontario, Waslander helped lead a team of students that converted a Lincoln MKZ sedan into a self-driving car dubbed “Autonomoose.”
The goal is to provide a “really nice, broad view” of self-driving car technology, including “how hard it is” to develop, Waslander said. Many people approach self-driving cars from specific related fields, such as robotics, and may have trouble seeing the big picture, Waslander said. Alternatively, students with less background can also use the program to choose an area to specialize in, which is what companies want to see, Waslander said.
Given the number of companies developing self-driving cars, supplying components like sensors, or working on related services, it’s easy to see how courses in self-driving car tech could be valuable. But despite the number of players, autonomous cars still face significant challenges.
“One of the remaining challenges is just corner cases,” University of Toronto Professor Jonathan Kelly said. These are “one-in-a-million scenarios” that are difficult for developers to anticipate, he explained. Waslander added that predicting the behavior of pedestrians is still a challenge. So is a scenario familiar in Canada, but not necessarily in environments like California and Arizona, where the majority of self-driving car testing takes place.
“They’re still not very good in cold weather,” Waslander said.
Updated on January 30, 2019: Corrected “four-week specialization” to “four-course specialization.”
- Lyft and Aptiv’s self-driving car program has come a long way (but not far enough)
- Bosch cautiously moving full-speed ahead with self-driving car tech
- Apple opens up about its self-driving car program in letter to NHTSA
- Waymo rules and Apple trails in California self-driving car benchmarks
- Sit back, relax, and enjoy a ride through the history of self-driving cars