Technology & Humanity
New Planning Method Could Make Driverless Electric Taxi Fleets More Efficient
From Phys.org Technology · October 21, 2026
Researchers at Chalmers University of Technology in Sweden have developed a new optimization method designed to improve the efficiency and reliability of driverless electric taxi fleets. According to the study, managing these fleets is complex due to shifting urban travel demands and the necessity of regular vehicle charging.
Sten Elling Tingstad Jacobsen, a researcher at the university, utilized Bayesian neural networks to predict travel demand based on historical data from San Francisco and Chicago. These networks provide estimates of uncertainty, which helps fleet management remain robust during unpredictable conditions like unusual weather. The method uses machine learning for forecasting, while a transparent optimization model makes decisions regarding vehicle placement and charging schedules.
The research, published in Transportation Research Part C: Emerging Technologies, aims to prevent urban traffic congestion caused by imbalances in vehicle availability. The approach breaks complex problems into smaller parts to remain computationally practical while scaling to large fleets.
A factual summary prepared by The Quantum Disciple. The original report belongs to its publisher.