Baidu Apollo World: AI simulator lets Apollo Go robotaxis practice up to 100,000 km a day
Baidu unveiled Apollo World, an AI system that builds a simulated city for training its Apollo Go robotaxis, at its AI Day in Shenzhen on Nov. 3. Baidu says it can generate up to 100,000 km of simulated driving a day, about three years of real-world testing by one safety-driver car.
Baidu says every car and pedestrian in the simulation acts as an AI agent that reacts to the robotaxi, and the system streams camera and lidar data in real time so the driving software can run complete trips, from pickup to drop-off. Baidu says it already uses it to train and test the Apollo Go fleet, drawing on more than 350 million km of real driving.
Why it matters
Robotaxi makers lean on simulation to cut the cost of real-road testing. Baidu has not said how well skills learned in Apollo World carry over to real streets.
More on robotaxis and self-driving vehicles.
The source post
Post by World in Motion on X.
View the post on X
Sources
- World in Motion (@WorldInMotionHQ) on X (aggregator)
- Reuters, via CNBC TV18: China's Baidu launches AI model simulating road environment for robotaxis
- Leiphone (Chinese): Baidu AI Day Apollo session
- Sina Finance / NetEase (Chinese)
New to robots? Read our guide: How robots work: a beginner's guide.
Summarized by Servospace from the sources above, drafted with AI help and checked against those sources before publishing. Claims are the companies' or outlets' unless we say otherwise. How we work.