Sterling Anderson: From MIT's Intelligent Co-Pilot to Aurora's Autonomous Future

Sterling Anderson, co-founder of Aurora, details his decade-long journey in autonomous vehicle development, beginning with his PhD work at MIT on shared human-machine control. He introduces the "intelligent co-pilot" concept, a system designed to enhance driving safety by modulating control between human and automation, ensuring vehicle stability within defined state-space constraints rather than enforcing rigid paths. This approach, which utilized techniques like Voronoi diagrams and Delaunay triangulation, aimed to prevent accidents by having the automated system intervene only when the vehicle approached its stability limits, then seamlessly returning control to the human. Anderson highlights the critical human interface challenges encountered, such as managing driver mental models and the counter-intuitive psychological effect where drivers reported feeling more in control even when the automation was significantly active.

The discussion transitions to Anderson's time leading Tesla's Autopilot program and the subsequent founding of Aurora. He emphasizes the significant shift in the automotive industry's recognition of self-driving, ride-sharing, and vehicle electrification as intertwined, transformative forces. Aurora's strategic business model focuses on developing the core self-driving technology stack and partnering with major automakers like Volkswagen Group and Hyundai. This collaborative approach is designed to achieve rapid, broad, and safe market deployment by leveraging partners' expertise in vehicle manufacturing and distribution, rather than building vehicles in-house. Anderson asserts that the economic case for operating self-driving fleets in mobility services is already viable, even with current sensor costs, advocating for initial deployment in high-end services before widespread cost reduction.

Anderson identifies forecasting the intent and future behaviors of other road actors as a primary technological bottleneck in achieving full autonomy, noting that this remains an unsolved problem requiring advanced machine learning techniques. He stresses that comprehensive capture of every possible driving scenario is an unbounded challenge, necessitating statistical evaluation to ensure a safety level superior to the average human driver before deployment. The conversation also delves into the profound societal implications of autonomous vehicles, including their potential to drastically increase safety, improve transportation access for underserved populations (elderly, disabled), and enhance urban efficiency by reducing the need for extensive parking infrastructure.

Finally, Anderson addresses the ethical considerations, particularly the inevitable job displacement in the transportation sector, underscoring the importance of proactive planning for transitioning affected workers into new roles. He envisions a future where personal vehicle ownership and driving evolve into a leisure activity or sport, while autonomous systems handle mundane or dangerous commutes. This paradigm shift could fundamentally reshape vehicle design, potentially reducing the reliance on extensive passive safety features like crumple zones and airbags in a world with significantly fewer collisions, opening up new opportunities for augmented reality and other in-car services.

"instead of designing in path space for the robot we instead found a way to identify plan optimize and design a controller subject to a set of constraints rather than paths"

"our biggest concern was if you throw off a human's mental model by causing the vehicles at behaviors to deviate from what they expect it to do in response to British control inputs that could be a problem"

"drivers reported feeling more control of the vehicle 12% more of the time when the copilot was engaged and when it wasn't and then noticed the statistics it turns out they actually at the average level of control the the copilot was taking was 43%"

"the automotive world has really come into the full-on realization that self-driving and particularly self-driving and ride-sharing and vehicle electrification are three vectors that will change the industry"

"our mission is to get a technology to market as quickly as broadly as safely as possible that mission is best served by playing our position and working well with others who can play theirs"

"one of the key challenges of self-driving rima is and remains that of forecasting the intent and B and future behaviors of other actors both in response to one another but also in response to your own decisions in motion"

"you will never have comprehensively captured every case every scenario that is as my some of you may want to correct me on this I think that's an unbounded set"

"I think it's incumbent on us to find a good way of transitioning those who are employed in some of the transportation sectors that will be affected into better work"

"presumably in a world where we don't crash there is there is much less need for passive safety systems"

"I think the opportunity really is to turn that it turned sort of personal vehicle ownership and driving into more of a sport and something you do for leisure"