Model based roll axis control of a terrain adapting Mars rover

This paper uses aspects of dead reckoning for "terrain adapting" a mobile robot moving over a rocky terrain like that of Mars. Besides the use of wheel encoders, an accelerometer and a rate gyro is used to sense the motion of the rover chassis and stabilize it, along the roll axis with the terrain adapting system. The rate sensor data is fused with the wheel velocity data from the wheel encoder and the accelerometer information to control the terrain-adapting arm. A contemporary Kalman filter system is utilized to condition the input to the model based system, and a nonlinear control for the robot\´s terrain adaptation system has been developed to generate asymptotic stability. Finally, simulation results have been developed considering the design requirements for an actual prototype.