Toyota may not be the first carmaker to investigate autonomous vehicles, but it hopes to set itself apart with a different approach to automated driving. The company is testing a self-driving prototype on public roads in Tokyo not only to achieve parity with other carmakers, but also to demonstrate its different philosophy.
Called the “Highway Teammate,” the car is a modified Lexus GS that represents the first practical step toward the use of artificial intelligence discussed by Toyota when it announced a major research collaboration with MIT and Stanford last month. Toyota says the technologies being tested could be production-ready by 2020.
Toyota says the car has already been tested on Tokyo’s Shuto Expressway in a series of trials covering functions like merging onto or exiting the highway, and maintaining or changing lanes. The driver can switch into automated mode only after passing through a toll gate and entering an on-ramp. The car uses detailed maps and its own sensors to orient itself, and keep track of everything around it.
While other carmakers and interests promoting self-driving cars sometimes emphasize removing humans from the loop, Toyota views automation as a partnership between people and machines. The company has hesitated to commit to fully-autonomous cars on a large scale, instead advocating intelligent systems to assist drivers, and fully automated vehicles for those who can’t drive, such as the elderly and the disabled.
To make this “Mobility Teammate Concept” possible, Toyota believes it will have to achieve a breakthrough in artificial intelligence. Up to this point, the majority of self-driving cars rely purely on programmed responses, making them only as competent as the people writing the software. Toyota thinks it can leapfrog these systems by developing cars that can actually think for themselves.
Specifically, Toyota is looking to develop three types of vehicular “intelligence.” One area involves the ability to recognize a situation and make decisions, another involves communication with other cars and infrastructure, and a third involves interactions with the driver, including recognizing the driver’s condition and coordinating when responsibilities should be shifted between human and machine.
- Good luck getting out of a ticket with Ford’s A.I.-enabled police car
- The U.K. is planning the ‘most complex journey yet’ for an autonomous car
- Ford’s self-driving cars are now delivering pizza in Miami
- GM says it’s serious about commercializing self-driving cars
- Waymo orders ‘thousands’ of Chrysler Pacifica minivans for its autonomous fleet