I recently attended a talk by Dr. Stefan Heck, on the future of urban transportation. I’ll shout out a thank-you to Acterra, the SF Bay area’s premier environmental stewardship organization, for hosting the talk.
Dr. Heck is a founder and CEO of Nauto, a Palo Alto, California startup that is combining deep-learning AI algorithms with vehicle dash-cams to learn safe, and unsafe, driving behavior directly from real-world drivers.
Beginning his talk, Dr. Heck quantified the tremendous costs and inefficiency of our current urban transportation infrastructure and practices. A typical personal vehicle sits unused 96% of the time, taking up space. When in use, only about 14% of the energy released by burning gasoline actually moves the vehicle, and most of that is to overcome the inertia of the vehicle itself, rather than the body and cargo of the typical lone occupant. All that energy, and all those climate-changing emissions, mostly to move a heavy hunk of metal, glass and plastic from one parking spot to another!
The indirect costs of personal vehicles are also staggering. Federally-subsidized multi-lane freeways are inefficiently utilized. The land-value of parking spaces drives up the costs of housing and commercial property. The United States loses $300 B in social costs due to accidents, and $300 B in lost productivity due to traffic congestion. The environmental impact and social costs are enormous.
Solving this problem is about policy as much as technology. Dr. Heck advocates making cities friendlier to pedestrians and cyclists, which went over very well with the Acterra crowd. Most of our urban areas need policies that encourage rather than discourage higher density urban centers (He’s talking to you, Palo Alto). Mid-rise buildings combining ground floor retail with office space and residences above, in the manner of many European cities, give efficient density in mid-size urban centers without the disruptive impact of high-rise towers. With residential, office and retail are concentrated, we could re-discover the physical and mental benefits of walking, and for trips up to 8 miles or so, we could ride electric bikes that fold up and go to the office with us.
The good news is that we are already seeing a dramatic decrease in car use by individuals under age 35. There are numerous factors contributing to this trend, including a preference for urban living among millennials, more electronic vs. physical social networking, greater environmental awareness, and openness to alternative modes of transportation. Broadly speaking, it appears that the American “love affair with the automobile” is gradually “aging out.” An automobile is increasingly seen strictly as a utility, rather than part of one’s identity.
The surprisingly rapid deployment and use of ride services like Lyft and Uber have made it far easier to go without a car in urban areas. If you own a car, you are more likely to use it, even when you could walk or bike. If you don’t own a car, and instead rely on ride services, you will want to use a ride service only when really needed. At less than about 4000 miles per year, it is actually cheaper (in the SF Bay area, at least) to use Lyft or Uber than to own a car, even excluding parking. Using on-demand carpooling can bring the break-even up to about 9000 miles, which is pretty close to what a typical urban driver puts on their car in a year.
Where policy and lifestyle changes prove resistant, technology can come to the rescue. Dr. Heck presented what he calls the ACES transformation: Autonomous, Connected, Electric, and Shared vehicles.
Sharing. Ride sharing (multiple passengers) will reduce the number of vehicles on the roads, while vehicle sharing (one vehicle used by many different people over the course of a day) will dramatically reduce the need for parking spaces, as well as the congestion and fuel waste caused by circling for parking places. Fewer cars on the road means lower road maintenance costs and no need to keep adding more lanes. Not mentioned by Dr. Heck, shared vehicles also have a faster payback and depreciation than personal vehicles, which allows earlier retirement and replacement by newer, better, and more efficient vehicles. This becomes ever more important as car technology changes more and more rapidly.
Electric. The energy-efficiency gain with electric vehicles is huge. The headline comparison given by Heck is that an electric motor is about 85% efficient compared to an internal combustion engine that is only about 30% efficient. A more accurate comparison would have to consider the greater weight of the electric vehicle, and the various energy losses from gasoline production and distribution versus those in electricity and battery production and distribution. Even with all that, the electric vehicle is probably about twice as energy efficient, and creates much less pollution and greenhouse gases. With battery technology advancing steadily, all-electric vehicles are becoming more cost-competitive. Even so, there has been resistance to adoption due to the inconveniences of charging the vehicle. This is a less important factor, however, with fleet vehicles, and will be even less of a factor once fully autonomous vehicles are deployed by fleet operators. Cars will simply head off to a charging station when needed, to be replaced on the road by a fully-charged vehicle. Battery-pack swapping will also be a great solution for fleets. Battery packs can be charged up using off-peak electricity, and quickly swapped into a vehicle to get it back on the road.
Connected and Autonomous. It is easier to take these together, because vehicle-to-vehicle and vehicle-to-cloud communications will be an essential part of all self-driving vehicles. Self-driving vehicles will not be alone on the road, nor autonomous in their actions. [See my article “The Autonomous Car that Won’t Be.”] Self-driving vehicles will share information, and act in coordination, to greatly improve safety and traffic flow. Already there are trials underway where vehicles form tightly-spaced “pelotons” that minimize energy loss to to air drag, reduce roadway congestion, and improve safety. For more on this, see my article “Autonomous Cars Will Run in Packs.” Partially and fully “autonomous” vehicles will also be connected to cloud services that will provide real-time traffic control and optimization. Dr. Heck’s startup, Nauto, looks to be an important player in cloud intelligence for connected vehicles.
In summary, ACES will completely transform our roadways, although it will take a while. A decade, perhaps. Nevertheless, I liked Dr. Hecks assertion that we have already or soon will hit “peak roadway,” and will before long have opportunities to remove rather than add lanes to our streets and freeways. Synergistically, that will open right-of-way space for pedestrians, bicycles, hop-on-hop-off electric buses, and rail systems.
Unfortunately, Dr. Heck did not have the time or opportunity to really discuss what his start-up company, Nauto, is up to. Initially, as I understand it, the Nauto systems will help fleet managers and insurance companies better understand the causes of traffic accidents. That understanding can then be the basis for driver training and for new driver-assist technologies.
It isn’t hard to see, however, that a thorough knowledge of human driving behaviors under a wide variety of conditions will be used to improve and inform self-driving vehicles. Direct observation of drivers and the road is an important third way of learning to drive, which can complement autonomous vehicle learning-by-doing in a controlled environment, as exemplified by Google, and radar-based mapping and vehicle sensor-based learning in real-world conditions, as exemplified by Tesla.
Having just won a $12 M Series A venture capital investment, Nauto is definitely a company to watch in the automated vehicle space.