Energy-Optimal Trajectory Planning with Vehicle's Dynamic Model Considerations

Indoor Autonomous Vehicles (IAVs) play a crucial role in load transportation, aligning with the objectives of Industry 4.0 in the industry sector. As IAVs operate on battery power, efficient energy usage is essential for mission completion. Hence, this paper presents a novel methodology for incorporating the dynamic vehicle model into the trajectory planning process to generate energy-optimal trajectories. To achieve this, the Timed Elastic Band (TEB) algorithm is selected as the base trajectory planner, which is then enhanced with a dynamic vehicle model. The approach is tested on an Autonomous Forklift (AF). The first step involves the development of a dynamic model for the AF, followed by the creation of a dataset using this model. Then a deep neural network (DNN) model is trained on this dataset to predict the required torques for the vehicle's straight-line and rotational movements. Finally, the trained model is integrated into the TEB planner, resulting in an optimized version, referred to as OTEB, for determining the most efficient trajectories. Simulation results demonstrate that the proposed trajectory planner significantly optimizes energy consumption, highlighting its effectiveness in real-world applications.