Regulations, Testing, Partnerships and the Hype around Autonomous Cars

First things first – Google or Alphabet has always maintained that its focus on self-driving cars where the driver has no direct involvement with the task of driving is to eliminate the human error from driving and thereby decrease accidents and fatalities. Conventional automakers on the other hand have voiced out the opinion that automated driving (as they prefer to call it) is an extension of active safety/ADAS systems to impact similar reductions in accidents and fatalities by reducing human stress and errors. One can argue at length in terms of which of these approaches will have the ultimate safety impact but the bottom line is getting these to the hands of customers is going to be one long marathon.

A few important announcements have been made public in the last few days/weeks that will add more fire to this discussion

• California DMV’s proposal to allow only automated vehicles with a driver behind the wheel with special licenses to be tested on public roads. The rulemaking which is still probably a year away has already witnessed sharp criticism from Google and praise from a few corners. How much of an impact the Tesla Autopilot beta feature had on this proposal is still unknown Tesla’s shouting on the need for the driver to be behind the wheel and attentive at all times.
• Coming on the back of Ford securing approval in California to road test its automated vehicle are speculations that Ford will be partnering with Google on its automated driving solution. While many are talking about this, let us not forget that the first car company that publically shared interest on working with Google was GM (https://www.carscoops.com/2015/10/google-and-gm-to-team-up-for-autonomous.html). And the best report I have seen on this news is from theverge that reports that the partnership could be in the form of an independent JV to protect each entity from liability and will also be a non-exclusive deal meaning Google can take this to other OEMs (https://www.theverge.com/2015/12/22/10648126/ford-google-self-driving-car-partnership-ces-2016). There are many indicators that make everyone believe that this could be true ranging from the early signals Google sent out that partnership with car companies is the way they are looking to commercialize its SDC software besides the ridesharing approach they are pursuing internally. Trust Ford to keep this under wraps till CES and if the announcement is true then the time to market for these solutions will definitely see a massive improvement.
• NHTSA proposal to revamp the NCAP star rating system in the US and most importantly adding a whole suite of ADAS features to the list to be implemented by MY2019. This is besides the commitment from 10 OEMs to install AEB systems in their vehicles in the coming years as a standard feature. The context - 8.1 percent increase in fatality rate in the first half of 2015 compared to first half of 2014. The context to this data is an increase of vehicles miles traveled by about 3.5 percent or 51.9 billion additional miles in first 6 months of 2015 compared to 2014.

 In terms of dissecting the possible impact of the California DMV proposed rule, there are a few interesting angles to pursue

• Could this possibly mean the balance shifting to Detroit? There are camps in the industry that believe that testing in a facility like Mcity in Detroit provides more real-world situations given the dynamic weather patterns, etc. OEMs like Ford have announced that they will soon start testing their vehicles in Mcity. And given that OEMs like GM have recently stepped up their internal test facilities for active safety adds more weight to OEMs using their own backyards for better results. But the flip side of testing in controlled environments is getting access to real world performance data like the Tesla Autopilot customer reactions and issues.
• Announcement from Elon Musk that further restrictions will be added to the Autopilot feature limiting availability and functionality. Already the Autopilot feature alerts the customer every 5 mins to place their hands on the steering wheel failing which the vehicle slows down and halts and doesn’t work on speeds above 90 Mph. The very fact that Tesla is doing this is testimony enough for adverse reactions from regulators and other OEMs. Known to be a disruptor in many ways, Musk and team took a lot of pain to emphasize the beta nature of the software and driver attention requirement but the crazy situations that many Model S drivers have gotten themselves into shows that too many complexities remain to be solved.
• Every little announcement around automated driving seems to be generating huge cycles of hype and reactions. Remember much of the industry is still in the development mode when it comes to automated driving right from choosing the right sensor combination, fine tuning that sensor fusion algorithm and let the system learn from the real-world conditions. This is where a massive amount of work is being done in the deep learning space from universities and vendors like NVIDIA. Road testing is not being ruled out but road testing of the Google vehicles are under the scanner. This is where concepts like workplace on wheels, lifestyle on wheels all need to get kicked out and the market needs to get back to reality. Of course Google still has a big play for its vehicles in a ride-sharing model in controlled environments like university campuses.