Add planes and trains to the self-driving vehicle revolution, which will go far beyond automobiles, a Google patent application suggests. Also throw in boats, snowmobiles, lawnmowers, golf carts, helicopters, earth movers, farm equipment and amusement park vehicles – Google’s list is long. And, according to the application filed in September and just posted by the U.S. patent office, “Other vehicles are possible as well.”
Those potential applications for self-driving technology were presented in Google’s application as examples of what’s possible for the technology it seeks to patent.
The application also provided rare insight into the value-judgments required of the artificial intelligence systems running autonomous vehicles, which must evaluate and act on the “risk magnitude” a particular driving action would pose. “Bad events may include various types of accidents, such as an accident with another vehicle, an accident with a police car or ambulance (for which the risk magnitude may be greater than it is for accidents with other vehicles), and/or an accident involving a pedestrian, among other possibilities,” the application said.
The document details the ways a robotic car would make decisions about positioning itself to maximize its awareness of its surroundings, indicating that the roadway decision-making by Google’s self-driving cars is based on risk-vs-benefit scores.
“The autonomous vehicle may evaluate information-improvement expectations and risk cost in the expected vehicle state corresponding to each possible active-sensing action,” the document said. “The autonomous vehicle may then determine a score for each active-sensing action by subtracting the respective risk cost from the respective information-improvement expectation. The autonomous vehicle could then perform the top-scoring active-sensing action.” Google did not explain in the application whether such scoring would in fact govern every move a robotic car made.
One scenario depicted in the application illuminates the relative badness of “bad events,” and the ethical decisions a robot car must make. If the car is stopped behind a large truck that is blocking its “view” of a traffic light, for example, the automated system must decide whether to move so it can “see” the light. Getting hit by the truck as a result of making the move warrants a score of risk-magnitude score of 5,000. Getting rear-ended is twice as bad as a side-swipe from the truck, bringing a 10,000 score. Twice as bad as the truck collision would be getting hit by an oncoming vehicle. But, according to the application, the robot car should consider one outcome to be five times worse, even, than a head-on, with a risk-magnitude of 100,000: “hitting pedestrian who runs into the middle of the road.”
Against resistance from the California Department of Motor Vehicles, which opposes fully automated cars, Google is pushing to put onto the roads self-driving vehicles that have no steering wheels or brake pedals. However, the patent application shows the company believes robotic cars must not make decisions that upset humans. “Bad events are not limited to accidents, and may include any event that could be perceived as negative in some way,” the application said.
“An autonomous vehicle may consider the dislike of an active-sensing action by a passenger in an autonomous vehicle to be a bad event (e.g., if an active-sensing action results in the autonomous vehicle repeatedly switching back and forth between lanes, or suddenly accelerating or coming to a stop). Similarly, the dislike or annoyance of passengers in other vehicles due to an active-sensing action may be considered a bad event.”
The patent application also describes the complex system of sensors the company is using for self-driving cars – the technology that allows such a vehicle to, essentially, see and hear everything around it, and orient itself. The sensor system, according to the application, can include cameras, microphones, GPS, gyroscopes, compasses, radar and lasers.
Photo: U.S. Transportation Secretary Anthony Foxx and Google executive chairman Eric Schmidt listen to Chris Urmson, director of Google’s self-driving car project, discuss the revolutionary vehicle at the Google campus in Mountain View last year (Karl Mondon/Staff)