Reprioritization of Autonomous Vehicles Development is Needed
Everyone is in the self-driving car race. Google has been developing autonomous driving capabilities since 2009 and continues to demonstrate incremental improvement in its autonomous driving technology since. By 2012, Google’s fleet has logged 300,000 self-driven miles. Tesla’s AutoPilot, another household name on the driverless car stage, has logged nearly 250 million miles of hands-free driving. Delphi’s Roadrunner completed a nearly 3,400-mile cross-country trip, 98% of it in self-driving mode.
Fearing the perception of falling behind in the race to take the checkered flag, practically all major automakers and suppliers are investing heavily in self-driving technologies and are very vocal telling their own bold vision of crash-free, stress-free traffic. And if you are a car executive who doesn’t dream big enough, you are out!
But despite truly impressive progress in self-driving technology these companies and others have made thus far, at the rate things are going, it will take a long time to get safe, reliable and commercially viable self-driving vehicles on main roads. And, in contrast to the excitement in Silicon Valley, Detroit and Stuttgart, customers don’t seem to care all that much. Most are skeptical about autonomous cars and are lukewarm about advanced semi-autonomous safety features.
Without sufficient consumer interest and adoption, progress will find few customers.
The SELF DRIVE Act that passed the House of Representatives in a unanimous vote is an important milestone in making self-driving vehicles a reality. Insurance policies that address uncertainties about accident liability are critical. And continued discourse about the ethics involved in automated decision-making is certainly prudent.
Several auto OEMs, and Tesla in the lead, say that guidelines aren’t enough to ensure self-driving technology is safe. They are pushing to get self-driving technology out on the road, and soon. They want to accumulate more real-world driving miles under a range of road and weather conditions to mature the technology and advertise their success to shareholders, transportation authorities, and the general public.
But critics express concerned about the maturity of the technology and its ability to handle real-world traffic conditions and human drivers, and point at several crashes of autonomous cars, some fatal, to cast doubt about the path these carmakers are taking. And as long as the road ahead includes a mix of cars driving by smart robots and others driven by stupid and impulsive humans, car crashes are likely to continue. Recently, Tesla engineers expressed hesitation about the readiness of the AutoPilot software and Elon Musk’s aggressive schedule to get it into production cars.
The road to reaching the ultimate goal of autonomous cars roaming the streets and cruising the highways is going to be bumpy and take longer than many expect. And accidents will surely happen. But these are not going change the direction in which OEMs are pursuing the vision. The deep-rooted automotive culture of head-to-head competition and prevalent one-upmanship is unlikely to change overnight.
Automakers should consider alternate routes, or, minimally, multiple parallel ones, to the current practice of setting prototype cars free and hope for the best. I am not advocating automakers should not pursue bold visions and aggressive and creative ways to mature autonomous driving capabilities. Rather, they should also use less glamorous but more attainable and practical paths to apply autonomous driving technology for tasks that are easier to understand, predict and mange. Examples:
- Autonomous driverless transportation to serve large hospital campuses, ports and airports, university campuses, large factories, industrial parks, and similar limited-scope applications. These areas are well-zoned and the likelihood of unexpected obstacle is very low. A variation of this application is limited-range neighborhood transit system such as those targeted by the Olli bus. At a recent Bloomberg Conference, Waymo CEO John Krafcik said the company may first commercialize its self-driving technology in the trucking and logistics sector, rather than the ride-hailing space This is the right business strategy, especially if the first vehicles are used for short-haul delivery applications rather than attempting to on the interstate.
- One application area begging for a change is the urgent need to replace the large, noisy, polluting buses that shuttle passengers between airport terminals and from there to multiple car rental locations, traveling nearly half-empty most of the time.
- Truck platooning systems, in which a human-driven truck is followed by driverless trailers. Platooning not only reduces the demand for experienced drivers. It also cuts congestion and reduces fuel consumption and emissions thanks to aerodynamic drafting. Early generations of this truck aren’t completely driverless: the lead truck is driven by a human driver, assisted by ADAS.
- Autonomous and semi-autonomous farming and agriculture equipment.
These and similar applications of autonomous driving technology aren’t entirely new. They are not radical either, and, admittedly, they do not offer the level of technology testing and stressing required to meet the end goal of safe and reliable autonomous passenger cars. But deploying driverless vehicles that the public actually uses will go a long way to demonstrate attainable value and improve consumers’ understanding and confidence in advanced safety and self-driving technology. Consequently, the growing adoption will further accelerate the maturation of the technology at a vastly lower risk and reduced cost.
Slowing down will help OEMs go faster.
And when you consider this development and market strategy, same approach should be taken by automakers to accelerate the adoption and lower the cost of electric vehicles.
Image: Time Warp (Mike Gambino, 2011)