Hierarchical Planning and Model Predictive Control for Passive Wheeled Terrestrial-aerial Bimodal Vehicles

To address the issues of modal discrepancies, ground nonholonomic constraints, and dynamic uncertainties encountered in the multimodal motion planning and unified control of passive-wheeled air-ground dual-modal robots, we propose an integrated planning-and-control framework. First, this framework comprises a hierarchical motion planner, including a front-end kinematic path search algorithm and a back-end gradient-based B-spline trajectory optimization algorithm. It generates a low-power, safe, smooth, and dynamically feasible land-air trajectory in unknown environments. It also includes a data-driven robust dual-modal unified controller and mode switching mechanism, enabling precise trajectory tracking control and stable mode switching. Finally, we validate the algorithm in both the ROS-Gazebo simulation platform and a physical platform, demonstrating the effectiveness of our approach.