Wheeled Mobility

Wheeled mobility research focuses on enhancing the capabilities of wheeled robots to navigate challenging terrains, particularly those with significant vertical variations. Current efforts concentrate on developing improved traversability estimation techniques, often employing data-driven approaches and reinforcement learning algorithms like Proximal Policy Optimization (PPO) to learn optimal control policies in complex environments. These advancements leverage techniques such as automatic curriculum learning and tight fusion of visual-inertial odometry with dynamic models to improve navigation accuracy and efficiency. This research is significant for expanding the operational capabilities of wheeled robots in off-road environments and has implications for autonomous vehicle navigation, robotics, and planetary exploration.