Data-efficient Iterative Learning Longitudinal Velocity Planning for Autonomous Logistics Vehicles

In recent years, the rapid growth of the logistics industry has brought rising fuel costs and increasing greenhouse gas emissions into sharper focus. With the development in autonomous driving technology, the problem of longitudinal velocity planning for logistics vehicles has attracted wide attention. To improve the fuel economy of autonomous logistics vehicles traveling on known routes, this paper proposes a data-efficient iterative learning optimization (DILO) method. The proposed method integrates the advantages of optimization-based and learning-based methods, using an iterative process to learn continuously from historical data and ultimately converge to the optimal longitudinal velocity solution. Simulation experiments conducted in Simulink and TruckSim using real-world road data demonstrate that the proposed method can generate optimized longitudinal velocity trajectories in both modes and significantly reduce vehicle fuel consumption.