Kinematic Control of a Planetary Exploration Rover over Rough Terrain

Passive averaging suspensions have been proven highly effective on rovers for improving mobility by providing ground compliance. However, the passive degree of freedom poses an added challenge to the controls problem. This paper presents a controller design to increase the accuracy of straight line trajectories for rovers with passive suspension on rough terrain. The chosen approach uses only proprioceptive sensors, a 3D kinematic model, and a trivial ground plane estimator algorithm to adjust individual wheel velocities based on estimates of terrain slope. This has distinct advantages to other techniques that use global position sensors and dynamic models which inevitably lead to more complex and computationally intensive solutions. The proposed controller is simulated in Matlab and found to be successful through experiments conducted with ORYX 2.0, a planetary rover research platform. This paper presents the feed forward velocity controller design, simulations, and experimental results for validation.