With a new grant, self-driving cars will be hitting the streets of UB’s North Campus to help researchers learn how best to bring autonomous vehicles into use by the masses.
Five years ago, Professor Chunming Qiao and the University at Buffalo began developing a three-in-one integrated driving simulator known as iCAVE2, short for Instrument for Connected and Autonomous Vehicle Evaluation and Experimentation. It’s a platform for evaluating self-driving vehicle technologies using three simulators – driving, traffic, and network – each with a different role.
“Driving simulator basically allows hardware and humans in the loop,” explained Qiao. “Traffic simulator allows us to produce the realistic traffic condition scenarios. Network simulator allows us to simulate wireless communication that’s needed for connected vehicles.”
Traditionally, researchers, automakers, and traffic engineers use simulators separately from real-world driving scenarios. A $1.2 million grant from the National Science Foundation will help Qiao and UB bring the two together by combining the current iCAVE2 simulator with real-world test-driving.
“It culminates our multi-year effort in this research area,” said Qiao.
“Its evidence that NSF and the research committee really think this is a worthwhile effort to pursue.”
UB will create a new facility for researchers to study how best to bring self-driving vehicles into mass use, and special equipment will be installed in and along the university’s roadways, turning the North Campus, itself, into a testing facility.
“We will be placing under-pavement sensors, cameras, what we call road-side equipment such as access points for wireless communication between those units and the connected vehicles themselves,” said Qiao.
Eventually, the “instrumented campus” will see a handful of “connected vehicles – equipped with wireless communication technology – hitting the roads. Since UB doesn’t have the budget for its own self-driving vehicles, partners at Carnegie Melon University and the Southwest Research Institute will bring their own to the Amherst campus for demonstration and experiments.
Qiao doesn’t expect regular vehicle drivers to see any great impact from the roadway research, although they will be part of the experiment.
“We need to study how human drivers or passengers react to those technologies,” said Qiao. “How other vehicles, for example, [interact] with connected or autonomous vehicles.”
Combining simulation with roadway testing also helps solve a disadvantage found in road testing on its own: safety.
“If you have unproven technologies, it might be unsafe for the researchers as well as for other travelers on the road,” said Qiao. He pointed out that self-driving technology also has limited flexibility for creating multiple ‘what if’ real-world scenarios, not to mention being costly and time-consuming.
In essence, UB’s research will “hit the sweet spot,” as Qiao put it, and bridge a crucial gap in the development of self-driving cars.
“There’s definitely a need for a third-party organization to do testing to make sure [automakers and IT companies] meet with whatever the standards will be for these vehicles, make sure they can inter-operate with each other, and they meet the safety, fuel efficiency, and whatever other requirements,” said Qiao.
Looking to the future, Qiao’s goal is for UB to lead the field in self-driving research, and become a national center for testing of connected and autonomous vehicle technologies.
