Numerical Stability of Inverse Simulation Algorithms Applied to Planetary Rover Navigation

Extending the navigational capability of planetary rovers is essential for increasing the scientific outputs from such exploratory missions. In this paper a navigation method based on Inverse Simulation is applied to a four wheel rover. The method calculates the required control inputs to achieve a desired, specified response. Here this is a desired trajectory defined as a series of waypoints. Inverse Simulation considers the complete system dynamics of the rover to calculate the control input using an iterative, numerical Newton - Raphson scheme. The paper provides an insight into the numerical parameters that affect the performance of the method. Also, the influence of varying the timestep and the convergence tolerance is examined in terms of the quality of the calculated control input and the resulting trajectory, as well as the execution time. From this analysis a set of parameters and recommendations to successfully apply Inverse Simulation to a rover is presented.