Toyota subsidiary Woven Planet will follow Tesla in using camera technology for its autonomous car project. Woven Planet said it can use low-cost cameras to collect data and train its self-driving system through a Neural Network, much like Tesla. This way, the company hopes to keep the costs low and scale up its self-driving vehicle project, a Reuters report quoting a company official said.
Tesla, for its autonomous vehicles, has been collecting data through cameras, and has found it to be a reliable and accurate way to determine the advantages and shortcomings of a system. Its Autopilot and Full Self-Driving suites are based on it. This approach is scalable, and less expensive than the autonomous vehicles that seek to collect data through sensors, which are decidedly costly.
Woven Planet works its way through
"We need a lot of data. And it's not sufficient to just have a small amount of data that can be collected from a small fleet of very expensive autonomous vehicles," Michael Benisch, vice president of Engineering at Woven Planet, was quoted as saying in the Reuters report.
The company plans to use small cheap cameras and artificial intelligence to gather data while driving and this will be the 'memory' for the autonomous car to plan its 'trips'.
But quite unlike Tesla, Woven Planet would not use just camera-alone technology. It will continue to use lidar and radar alongside cameras for collating info.
For its autonomous vehicles project, Woven Planet has made a few strategic acquisitions. It recently acquired the company Level 5, which is a division of Lyft, formed in 2017, specifically dedicated to self-driving technologies. In four years, the self-driving division has reached public road testing of its fourth-generation platform.
Woven Planet has also taken over Carmera Inc, a US-based spatial AI company specialising in next-generation road intelligence in automated mobility. Carmera, a barter-model based company, uses data collected from a service it provides for free to commercial fleet operators to maintain and expand its primary mapping product. Carmera’s primary offering is a high-definition map developed for autonomous vehicle customers.
By partnering with software stack experts like Apex.AI and absorbing self-driving technologies from Lyft and Carmera, Toyota through Woven Planet is trying to get on the top of the self-driving vehicles game. Woven Planet already has satellite-based mapping and the massive amounts of data gleaned from its millions of vehicles on the road today.
