Self-driving vehicles are tearing up the streets of cities as never before. In August, the California Public Utilities Commission (CPUC) granted two companies, Cruise and Waymo, the right to operate fleets of driverless robo-taxis throughout the day within San Francisco and to charge passengers for the transportation services. This was the latest of a string of green light that has granted more freedom to autonomous vehicle (AVs) within San Francisco in recent years.
Soon after, widely-publicized reports surfaced of Cruise vehicles acting strangely. One was a roadblock in front of a major music festival, a third got caught in wet cement and one even hit an ambulance. These accidents have brought a fresh focus to the challenges of integrating autonomous vehicles in the urban setting. In less than ten days following the CPUC’s approval at a 24/7 rate, The California Department of Motor Vehicles demanded that Cruise reduce the size of its fleet operating within San Francisco while the regulator investigates the accident involving a firetruck.
Despite these issues and ongoing protests, autonomous car companies have been able to continue expanding the scope of testing as well as operations in greater U.S. cities. A majority of those who support them believe that these cars could provide an option for human motorists, improve transportation more accessible to people, enhance traffic flow within cities, and reduce the environmental impact of vehicles. However, they have not yet delivered on their promises; the vast majority of those benefits are still in the potential future. However, now, the pitfalls have been made clear. Scientific American asked experts from the field what’s essential to know about the present state of the rollout of robot-taxis.
HOW DO ROBO TAXIS WORK?
AVs, including the ones operated via Cruise and Waymo, should not be confused with commercially accessible automobiles, such as Teslas, which come with certain driver assistance features. In contrast to those cars that need an individual driver to supervise the vehicle at all times and at all times, robo taxis exhibit what’s commonly referred to as “Level 4” autonomy. They are able to move through a specified space without having a person watching and controlling every move via the wheel. This is possible thanks to an extremely precise internal map, according to Ramanarayan Vashudevan, an engineer studying self-driving vehicles and is a professor at the University of Michigan.
This satellite-linked map covers only the area of a certain size within which the automobiles have been “geofenced,” meaning they aren’t allowed to operate beyond their boundaries. “The map is like a cheat code,” Vasudevan says. It is the reason why vehicles can navigate through complex terrain. They also depend on an extensive array of sensors, which includes sensors for sound, cameras and laser-based lidars that all aid in locating a car within the lanes of the map and provide real-time data on conditions like weather, pedestrians passing by construction, and other road conditions.
Machine-learning algorithms that have been trained using large amounts of both modeled and actual driving data–translate a car’s sensor and map information into projections of what bicycles, cyclists or cars will most likely to be doing, Vasudevan says. Artificial intelligence software also instructs the vehicle how to move and how to move, while pre-set guidelines ensure it is following traffic rules. In the event that it becomes stuck in an awkward circumstance, says Waymo the software engineer Josh Herbach, a human operator can direct it remotely to carry out the right step from a desk in the control center.
ARE ROBO TAXIS SAFER THAN HUMAN DRIVERS?
Waymo and Cruise maintain that their AVs are more secure as human motorists. The two companies have published information which they claim prove this and some independent researchers are in agreement. However, the reality may not be as simple as their statistics suggest, according to Steven Shladover, an engineer at the University of California, Berkeley who has been researching the automation of vehicles for more than 50 years.
“I actually think there is not yet enough data available to the public, including researchers like me, to be able to judge” the relative risk, Shladover says. Companies that provide AV services don’t share all of the data they collect. In most cases, incidents that are negative only become publicized when they are recorded or shared through social media. Even the information about human drivers aren’t always complete because they are limited to major accidents, Shladover adds.
The data released could be flawed since in order to make a precise comparison of robot taxis and humans in the same way, you must consider how many miles driven, and under what conditions – data which isn’t even available. There are more human accidents than vehicles do however more than 100 million human drivers are believed to be driving through U.S. roads every day as compared to just a few hundred of taxi drivers. In analyzing estimates of the number of driving hours drivers spend on the road, “crashes with human drivers are actually extremely rare events” in relation to the total amount of time that humans drive, Shladover says. Statistics obtained from National Highway Traffic Safety Administration appear to support this assertion..
While a perfect, automated system will certainly be more efficient than a human being, Vasudevan notes that such an automated system isn’t even possible. “Our ability as humans to build millions and millions of lines of code in a way that is 100 percent correct is nearly impossible for the same reason that humans are imperfect at driving,” Vasudevan says. If asked if he believes AVs are more secure drivers than human drivers, Vasudevan generally agrees with Shladover. “I don’t think that story is as clear as self-driving car enthusiasts would like us to believe,” Vasudevan declares.
WHAT ABOUT OTHER BENEFITS?
Beyond safety, a widely claimed benefit of AVs is that driverless vehicles will increase accessibility to transportation and reduce costs and reduce the requirement for owning a car. Another benefit is that the presence of more AVs are on the road could help reduce congestion in traffic and simplify urban planning by allowing cars to talk with one with each other and making sure that they are following the most efficient routes at the appropriate time. But neither of these scenarios is yet a reality, and the likelihood of achieving them isn’t 100 according to Dan Chatman, an urban planner and researcher in transit at U.C. Berkeley.
While he’s optimistic about AVs and is excited about its potential for San Francisco, Chatman acknowledges there’s a possible future where robo taxis create more problems than they solve. They can contribute to urban sprawl, cause congestion on roads instead of alleviating congestion, and deter users from the most environmentally sustainable modes of transport: walking, biking as well as public transit. “A lot depends on the extent to which they are regulated,” he declares. There are some policies that could help people ditch their automobiles, like road pricing, which will ensure that the real costs of environmental and social driving a car, regardless of whether it is human or automated-driven–is borne by its driver. “We haven’t done an excellent job of controlling automobility in general. So this is a chance to rethink how we interact with” automobiles, Chatman says.
There are two possibilities ahead of us: one that has proper regulation, care and transparency, and the other that doesn’t have them. Robot taxis could be operating on their own however it’s our responsibility as humans to decide which one we’ll follow.