Guidance and control of an autonomous planetary rover

This thesis presents a set of algorithms for piloting an autonomous planetary rover along a planned path, performing real-time obstacle avoidance and improving the dead reckoning capability. Path-tracking is accomplished through linear regulation (feedback) of position and orientation errors, measured with respect to the planned path trajectory. Obstacle avoidance is accomplished by applying the concept of an artificial potential field to data that may be acquired using a scanning rangefinder. Dead reckoning is accomplished through the algorithmic fusion of odometry and inertial navigation data. Results from computer simulation illustrate the path-tracking and obstacle-avoidance capabilities, where the system model is that of a three-axle rover, with symmetric front and rear-axle wagon steering. Experimental data, acquired using a vehicle fitted with a wheel-encoder and an accelerometer, illustrates how the fusion algorithm mitigates the effects of wheel-slippage and integration-error.