Efficient Attitude Estimation for Mobile Wheeled Robots Using Elevation Maps

Ensuring the dynamic stability of a wheeled robot in off-road navigation requires continuous state monitoring to anticipate risks such as slipping, skidding, or overturning. While proprioceptive sensors can monitor the robot’s state, fully predicting it remains challenging since its variables depend on robot-terrain interaction. This work aims to propose a direct geometric method to estimate the robot’s attitude by approximating wheel-ground contact with circles and leveraging an elevation map. Our approach is intuitive, straightforward to implement, precise, and computationally efficient. Experimental results, both in simulation and using a real robotic platform, validate its effectiveness.