People are funny. We’ve complained about having to waste time sitting uncomfortably in traffic for decades now. But when the phrase “self-driving car” and the idea of traveling in a car without having to dedicate full attention to it started becoming unavoidable in auto news, drivers of all sorts cried foul, calling the idea bad for reasons ranging from practical and real to theoretical and imagined.
Too far along to abandon the self-driving idea, automakers experimented with new language; disruptor Elon Musk demonstrated his wisdom with words by naming Tesla’s system Autopilot, after an established technology that’s already trusted and relatively understood, at least conceptually. Another important differentiator for Tesla is the fact that Autopilot promises partial rather than full autonomy, a critical difference that came up repeatedly at the recent New England Motor Press Association (NEMPA) and Massachusetts Institute of Technology (MIT) conference on The Intersection of Technology and Design.
The conference was hosted in MIT’s Media Lab by Paul Parravano, co-director of the school’s office of government and community relations, and featured a number of esteemed guests: Timothy Anness, Fiat Chrysler Automobiles’ Head of Advanced Design; Michelle Finamore, Curator of Fashion Arts at the Museum of Fine Arts; Mary Gustanski, Vice President of Engineering and Program Management for Delphi, a global supplier of auto technology; John Leonard, MIT professor and Associate Department Head for Research in Mechanical Engineering; and Gil Pratt, CEO of the Toyota Research Institute. This year’s event marked the sixth year of the conference and was supported by the companies that sent panelists, as well as Chevrolet, Mazda, and Mercedes-Benz. It also attracted a larger and more diverse audience than ever before.
The conference kicked off with a lunch during which Eric Wingfield discussed the new FordPass “mobility solutions” app, a system that will allow users to access and utilize multiple transportation modes seamlessly, which Forbes credits with turning Ford into a car and mobility company and Mashable calls “the beginning of the end of car ownership.” Ford plans to launch the new app and system with the 2017 Escape, and we look forward to seeing what the general public makes of it.
After lunch, Timothy Anness showed slides of some very futuristic yet still classic-looking concepts his team has produced and quoted Bill Gates as having said, “We always overestimate the changes coming in the next two years and underestimate the amount of change in the next 10 years.” That quote resonated with other panelists and came up again multiple times during the rest of the presentations. Anness was followed by Michelle Finamore, who discussed the many influences fashion has had on automotive design over the years, including 3D printing and paints that respond to changes in their environment.
Mary Gustanski noted that Delphi is evolving from an auto-supply company to a technology company as a result of the fact that future cars will include four times as many lines of code as an F-35 fighter jet. She noted that while Delphi is working on Level 4 fully autonomous driving systems, she thinks those will require more time, and she and the company are more excited about current and developing Level 2 and 3 systems that incorporate active safety and driver-assistance features like blind-spot monitoring, cross-traffic alerts, and automatic emergency braking while leaving the driver in primary control of the vehicle.
MIT professor John Leonard gave an entertaining presentation of his work on navigation and mapping for autonomous robots. Showcasing Duckietown, a model city packed with roadways designed to test and demonstrate small robotic cars his team members built for $150 each, Leonard noted that while most driving is “routine,” any fully autonomous Level 4 system will have to enable vehicles to deal with unexpected and potentially confusing events, such as a policeman waving a driver through a red light. While the ability to handle those sorts of events will be required very rarely, figuring out how to give robots that ability is one of the biggest challenges facing the folks working on autonomous cars.
Toyota Research Institute’s Gil Pratt gave the last presentation of the afternoon and focused on the differences between two potential solutions for future cars, “The Chauffeur,” which would be a Level 4 or 5 self-driving vehicle that will take care of everything, and “The Guardian Angel,” a less-sophisticated system consisting primarily of active safety systems like those we’re currently seeing in a wide variety of new cars, leaving the driver in control of—and liable for—the vehicle’s operation most of the time.
Pratt noted that because humans drive very well statistically, with only about 2 crashes per million miles of driving in the U.S., a Chauffeur approach would have to be fantastic to be considered an improvement over a human driver. That approach would also lead to human drivers becoming less competent over time and require manufacturers to shoulder liability for accidents. The Guardian Angel approach would be substantially less disruptive, leaving drivers in control of the car most of the time and liable for most accidents. Those differences suggest that Tesla’s decision to pursue a more Guardian Angel-ish approach, at least for the short term, was probably a wise one.
As the variety and depth of these presentations, not to mention the winners of NEMPA’s 2016 Winter Vehicle Testing, which were presented after the panel discussion, made clear, the future of self-driving cars has already arrived, and while technologists are focusing on a variety of approaches and tools, we’re confident we’ll be very surprised by the progress made over the next 10 years. But we’re not sure Bill Gates is right in this case, as we suspect we’ll also be surprised by the progress made in this realm in the next two years.