Shenzhen’s autonomous public transportation

Shenzhen demonstrates how AI can improve public transport well before fully driverless vehicles become universal. This U4SSC case study examines a privacy-preserving passenger origin-destination system developed by Intellifusion and Shenzhen Bus Group. Instead of facial recognition, the system analyses head, shoulder and neck features to reconstruct passenger flows and generate boarding and alighting data. These insights support six operational functions: designing express routes, recommending short-turn services, configuring mixed vehicle fleets, simulating route cancellations, identifying and correcting network blind spots, and coordinating bus and metro services to remove duplication. Reported outcomes include a 20 per cent improvement in dispatch efficiency, a 30 per cent reduction in operating costs, real-time occupancy monitoring and ten express routes that reduced travel time by 21 per cent for 70 per cent of passengers. The case also cautions against stale data, algorithmic errors and over-reliance on automated recommendations. More frequent data updates, citizen feedback and continued human judgement are therefore essential to make AI-led transit optimisation both efficient and responsive to users.

More information available at: https://u4ssc.itu.int/