The autonomous car world has buzzed with the recent announcements of crashes and a fatality related to the semi-autonomous Tesla Model S. This comes at a time when regulatory efforts surrounding autonomous vehicle operations in the U.S. are pushing decisions from Washington D.C. over standardization of technology to encourage faster evolution of autonomous cars.
Last week, I posted a blog on this topic. The concern within the industry is that the recent crashes will derail the overall development of autonomous vehicles, or at the very least significant impede progress. It is very tragic to hear of a life lost from an Autopilot operation, and my heart goes out to the family of the driver. However, I sincerely hope that the industry continues to push for safer alternatives to human driving, whether this will lead to autonomous passenger cars in the next decade or not until after 2025. Having driven over 200 miles this weekend for leisurely pursuits, I am struck by the increasingly risky behaviors I observe on the interstates around me (Disclaimer: I live in Massachusetts, which may skew my perceptions, given our propensity for aggressive driving). Numerous times I observed cars exceeding the speed limit by 20+ MPH, swerving in and out of lanes without using signals, and twice I was almost side-swiped when a car sped into my lane, seemingly from nowhere.
Now, this last example is something that can already be solved with some of the semi-autonomous technology that is available in many new cars today - lane change assist. In addition, there are a number of other excellent technologies that in essence move today's connected car toward autonomy, known as advanced driver assistance systems (ADAS). In fact, a consortium of the leading automotive OEMs worked together to make auto-braking, a category of ADAS, mandatory in all new U.S. vehicles by 2022. Another recent blog expands on this topic.
How should the industry work together with national regulators to ensure that autonomous vehicles continue their development while at the same time improve safety on the roads? I believe that the owners of autonomous technologies (automotive OEM's, major components suppliers, technology companies that are developing autonomous vehicles) need to share their data around how the cars are interacting in their environments. What was the Model S "thinking" when that fatal crash occured? What type of decisions is a semi-autonomous car making when it loses sight of road lines, or can't identify an obstacle?
My colleague Ruthbea Yesner Clarke, who leads the global Smart Cities research at IDC Government Insights, and I recently co-published a report on the collaboration needed between automotive OEMs and city planners to implement connected car and smart city solutions. Subscribers can access the full report here.
I understand these different players working toward creating the autonomous car of the future are competitors, and that there are trade secrets in the way each is developing the cognitive systems that operate the car. However, if there is not some form of data and experience sharing around the autonomous vehicle behavior in the real world, the further development of the technology will take years longer before it can be trusted. And, it is highly likely that during this time the general public sentiment may swing decidedly toward NOT being fans of autonomous vehicles. Then, the industry will have a much bigger problem on its hands.