In situ skid estimation for mobile robots in outdoor environments

Skidding is a surface hazard for mobile robots' navigation and traction control systems when operating in outdoor environments and uneven terrains due to the wheel-terrain interaction. It could lead to large trajectory tracking errors, losing the robot's controllability, and mission failure occurring. Despite research in this field, the development of a real-world feasible in situ skid estimation system with the capability of operating in harsh and unforeseen environments using low-cost/power and ease of integrating sensors is still an open problem in terramechanics. This paper presents a novel velocity-based definition for skidding that enables a real-world feasible estimation for the mobile robot's undesired skidding at the vehicle-level in outdoor environments. The proposed technique estimates the undesired skidding using a combination of two proprioceptive sensors (e.g., inertial measurement unit and wheel encoder) and deep learning. The practicality of a velocity-based definition and the performance of the proposed undesired skid estimation technique are evaluated experimentally in various outdoor terrains. The results show that the proposed technique performs with less than 11.79mm/s mean absolute error and estimates the direction of undesired skidding with approximately 98% accuracy.