Companies testing self-driving cars, like Google, love to boast that the new technology is already safer than human driving. When the statistics bear out the claim, it’s commonly assumed that machines cause fewer accidents owing to caution and the lack of human recklessness instead of agility and driving prowess.
But experimental designs can already make daring maneuvers outside of the capability of most drivers, and it’s not difficult to imagine a future where machine driving outpaces human steering in all respects.
On Oct. 20, Stanford University unveiled a modified 1981 DeLorean that could drift, a flashy move usually found in professional motorsport, without human guidance, making doughnut-shaped skid-marks across an empty lot. Drifting is when a car makes a turn so sharp that the back-wheels of the vehicle skids in the opposite direction of where the front-wheels are pointing.
The engineers that designed the car were motivated by the need for autonomous vehicles to have access to the full range of driving maneuvers that were capable of preventing car accidents, including maneuvers that are impossible to perform on most vehicles due to the prevalence of electronic stability control (ESC), which automatically applies brakes to a car when it’s skidding.
The ESC safety mechanism, which was popularized in the mid-1980s, has the drawback of limiting cars from making extremely sharp turns, which often result in skidding.
The researchers had another reason for choosing the 1981 DeLorean besides the lack of an ESC. The car is the same model featured in the 1985 cult classic “Back to the Future,” and is named MARTY (Multiple Actuator Research Test bed for Yaw), after the film’s main character, played by Michael J. Fox.
The Stanford News piece came out the day before Oct. 21, 2015, known as Back to the Future day, the day Marty McFly travels to in the future to save his yet to born children in “Back to the Future II.”
“In our work developing autonomous driving algorithms, we’ve found that sometimes you need to sacrifice stability to turn sharply and avoid accidents,” Chris Gerdes, a professor of mechanical engineering who designed the car, told Stanford News. “The very best rally car drivers do this all this time, sacrificing stability so they can use all of the car’s capabilities to avoid obstacles and negotiate tight turns at speed.”
In the next stage of testing, MARTY will be programmed to drive along a racetrack and drift to avoid obstacles, instead of just circling around in an empty parking lot. The ultimate goal of the project is to build up autonomous driving capabilities so that machines can match, or better, the best professional race-drivers, and not just rhetorically.
Eventually, Gerdes wants autonomous cars to participate in drifting contests with professional race-drivers, which he says is an excellent proxy for both the precision of autonomous car control and how it will interact with humans, as the contest requires careful maneuvering for the cars to avoid collision.
“While we aren’t picturing a future where every car produces clouds of white tire smoke during the daily commute, we do want automated vehicles that can decipher the subtle cues drivers give when driving and incorporate this feedback when planning motion. Drifting is a way to study these larger questions, with style,” Gerdes said.
