Toyota wants to develop self-driving cars with cameras instead of lidar

Toyota’s Woven Planet division is developing a system for autonomous cars without expensive sensors, such as lidar. Instead, the technology should use cameras, as Tesla does.

Woven Planet, a Toyota subsidiary that develops new technology, tells Reuters news agency that it can use relatively inexpensive cameras to collect data on things like traffic situations. This data can then be used to develop a system for autonomous driving. The company says that such data is crucial to developing such systems, but that it is too expensive and unscalable to collect data from a fleet of cars with more expensive lidar sensors.

The company claims the new camera system is up to 90 percent cheaper than the sensors it used before, saying it can be installed relatively easily in large numbers of passenger cars.

Tesla uses a similar method, in which the company collects data based on camera images from Tesla vehicles. Tesla’s Full Self-Driving option is currently offered on a subscription basis, but is still in beta. CEO Elon Musk said Tesla could achieve fully autonomous driving with cameras this year. However, the company has previously missed deadlines for this that were predicted by Musk.

The Toyota division does indicate that it will still release self-driving cars with lidar and other sensors for the time being, since according to the company this is the best and safest way to deploy robot taxis and other self-driving cars on the road. Woven Planet indicates that autonomous systems based on cameras can catch up in the longer term, but says that it does not know how long that would take.