Apollo Go robotaxi pilot — driverless cars with a safety officer still on board
Baidu's Apollo Go (萝卜快跑) robotaxi service began carrying passengers in Chengdu's Hi-Tech Industrial Development Zone in July 2022, initially with 8 vehicles across a 10km² area; by mid-2024 the operating zone had grown to roughly 60km² of roads through the Xinchuan Innovation Science Park, covering commercial, residential and tech-park areas, but the fleet was still just 8 cars running 8am–6pm with a paid safety officer in every seat. Demand outstripped supply: peak-hour waits of nearly two hours were reported in July 2024, and each car reportedly handled roughly 300 orders a month. Chengdu's operator says the service ran without a reported traffic accident since it began in 2022, but — unlike Baidu's fully driverless services in Wuhan, Shenzhen or Beijing — Chengdu's deployment had not gone safety-driver-free as of the most recent reporting found.
📊 Impact
8 vehicles covering ~60km² of roads (up from an initial 10km²) as of mid-2024; ~300 orders per vehicle per month; peak-hour waits of up to ~2 hours reported; no reported traffic accident since the 2022 launch (operator-reported, press-covered).
🎓 Lesson
A robotaxi headline ('driverless since 2022') can coexist with a fleet too small to matter — Chengdu's 8 cars and human safety officers show the gap between a permit and a service.
📖 Terms in this entry: Adaptive traffic signals · CBTC (communications-based train control) · Digital twin
Sources
This overview was generated with AI from the public sources listed above and checked against them before publication. It has not had a complete human fact-check — treat it as a starting point and follow the sources.
Continue exploring
Compare this project
Explore the wider context
📎 Cite this project
1001 Smart Cities (2026). “Apollo Go robotaxi pilot — driverless cars with a safety officer still on board” — Chengdu, China. The Smart City Atlas. https://1001smartcities.org/projects/chengdu-apollo-go-robotaxi/ (last verified 2026-08-02). Data: CC BY 4.0.
The underlying data is free to reuse with attribution — see the open data page.