BEIJING: Baidu launched an artificial intelligence system on Tuesday to simulate real road settings and train robotaxis. The new tool aims to help driverless car firms cut high tech costs.
Named the Baidu Apollo World AI model, the system builds up to 100,000 kilometres of simulated driving daily. That equals about three years of real road tests with a safety driver.
The software creates live camera and LiDAR data feeds. These feeds let self-driving cars see and navigate virtual traffic.
Baidu is already using the tool to train its Apollo Go robotaxi fleet for global growth. As China’s top robotaxi firm, Baidu faces local rivalry from firms like WeRide, Pony.ai, and Momenta Global.
Read: Apple Partners with Baidu for AI Deployment in China
Driverless car firms often supply software to local partners to avoid buying large vehicle fleets. Yet training these AI models remains very costly.
Read: Cars and speakers: Baidu speeds up AI progress
Financial firm Morgan Stanley projects that China’s robotaxi fleet could hit one million cars by 2035. Baidu launched Apollo World, an artificial intelligence system that simulates road environments to train robotaxis and cut tech costs.
(With input from Reuters)
