Spatiotemporal Prediction-Driven Model Predictive Control for Vehicle–Aircraft Conflict Resolution on Airport Surface

The increasing density and complexity of airport surface operations have intensified the risk of crossing conflicts between ground service vehicles and taxiing aircraft. Such interactions are characterized by strong spatiotemporal coupling, asymmetric right-of-way relationships, and stringent safety requirements, making conventional human-driven conflict avoidance highly dependent on drivers’ perception and judgment. To address this problem, this study proposes a spatiotemporal prediction-driven model predictive control (MPC) framework for autonomous ground vehicles on airport surfaces. First, the spatial interaction between the aircraft safety boundary and the vehicle service road is modeled to define the vehicle–aircraft conflict zone. Aircraft motion information is then used to predict the temporal occupancy of the conflict zone, based on which a dynamic time-window constraint is constructed to characterize the time-varying safe passage conditions for autonomous vehicles. The predicted spatiotemporal constraints are embedded into a rolling MPC framework that continuously optimizes vehicle motion while jointly considering safety, traffic efficiency, and energy consumption. Simulation results show that, compared with human-driven vehicles, the proposed method reduces average energy consumption from 1200.16 kJ to 866.67 kJ and shortens average arrival time from 43.63 s to 41.27 s. In addition, the method demonstrates effective disturbance compensation under aircraft-state uncertainty and adaptability to sequential multi-aircraft crossing scenarios.