An intelligent transportation system for autonomous bus transit under traffic uncertainty

This paper introduces ABUST (Autonomous BUS Transit), an intelligent transportation system designed to support demand-responsive public transport through autonomous buses in urban environments, with particular emphasis on improving accessibility for passengers who experience difficulties using conventional fixed-route bus services. The proposed system integrates routing , scheduling , and motion planning within a unified hierarchical architecture supported by fuzzy logic models. Passenger requests are modeled as a dial-a-ride problem, incorporating pickup and delivery constraints together with temporal requirements. Uncertainty in urban traffic conditions is addressed through a fuzzy inference model that modifies travel distances based on qualitative traffic states, while a second fuzzy model represents passenger satisfaction with respect to waiting and travel times using soft time windows. The resulting optimization problem is solved using genetic algorithms, enabling efficient treatment of the underlying computationally intractable routing and scheduling tasks. At the operational level, a robot-based motion planning model generates feasible speed and steering profiles consistent with vehicle kinematic constraints to ensure safe and comfortable bus operation. The proposed framework is evaluated through simulation experiments conducted on the urban road network of Athens, Greece. The results demonstrate that ABUST can effectively adapt to varying traffic conditions while balancing service efficiency, passenger satisfaction, and motion feasibility, supporting its potential as a flexible approach for autonomous, demand-responsive bus transit in smart cities.