Wheels: Semiautonomous Cadillacs Hit the Road for Testing

Monica Almeida/The New York Times Cadillac says its Super Cruise semiautonomous technology can allow motorists to let the car’s computer take over in bumper-to-bumper traffic typical of cities like Los Angeles.

It seems inevitable that there will be more and more cars with a computer rather than a human being behind the wheel in the years to come, and Cadillac has joined the race for auto autonomy. The company said Monday that it had begun testing a semiautonomous car, a vehicle that can handle driving functions without operator input in certain situations.

Cadillac calls its system Super Cruise, and it combines lane centering technology and full-speed adaptive cruise control to allow the automobile to handle steering, braking and speed regulation in situations like bumper-to-bumper traffic and highway driving, where direct driver input can be tedious and, in theory, unnecessary. The system would use components and technology that are already available in some cars, including radar, cameras, proximity sensors, and GPS data.

Dan Flores, a General Motors spokesman, said in a telephone interview that while Super Cruise would depend largely upon existing hardware, a lot of other technology was required as well.

“It involves taking in all this data, fusing it together and deciding what to do with it,” he said.

G.M. first introduced Super Cruise at the Association for Unmanned Vehicle Systems International’s Driverless Car Summit in Detroit last June and has said the technology could make its way into production cars later this decade. The system has progressed beyond laboratory simulations and is being tested on the road.

“The best way to achieve reliable performance is to gather as much data as possible in the conditions our customers will experience,” said Jeremy Salinger, the research and development manager for Super Cruise, in a media release.

Cadillac says that when Super Cruise handles mundane driving chores, operator attention will still be required, adding, “When reliable data is not available, such as when there are no lane markings, the system will prompt the driver to resume steering.”

Asked how the car would be able to determine whether drivers were paying attention before requiring them to take over, Mr. Flores said, “That’s an issue we’re working on right now. We believe one of the challenges to semiautonomous driving is how do you determine that the driver is engaged.”

Bryan Reimer, a scientist studying driver behavior at the Massachusetts Institute of Technology, has expressed reservations about semiautonomous technology, saying that it’s difficult to keep drivers involved if they have nothing to do. He said taking the workload away from a driver could cause understimulation and that was being lost in the debate about the future of part- and full-time driverless cars.

A basic tenet common to many studies of driver performance is the Yerkes-Dodson Law, a theory developed in the early 20th century that says human operators fail when there is too little to do and when there is too much to do.

Asked if the Super Cruise system might include a workload detection component that would cause it to take over control if the driver were overworked, showing signs of stress, Mr. Flores said, “There are a lot of things we’re looking at.”

“The intention is to bring Super Cruise to market later this decade,” he added, “but there are a lot of details we have to resolve.”