Human in the Loop Automation: Ride-Hailing with Remote (Tele-)Drivers.

This article focuses on tele-driving, a concept where drivers remotely operate vehicles, and analyzes how having fewer remote drivers than vehicles affects ride-hailing system performance. Using a spatial queueing model with state-dependent service times, the study identifies three operational regimes—supply-limited, intermediate, and supply-rich—based on vehicle capacity relative to workload. Key findings include that in supply-limited and intermediate regimes, reducing the number of remote drivers can paradoxically improve service levels or stabilize otherwise unstable systems by shortening pickup times due to more idle vehicles. In the supply-rich regime, significant reductions in driver numbers are possible without substantially degrading service quality. The analysis extends to both impatient customers (who leave if not immediately served) and patient customers (who wait), with numerical experiments based on New York City taxi data supporting the theoretical results. The paper also discusses implications for dispatch policies and highlights broader applications of tele-driving beyond ride hailing.