An improved formulation for optimizing rocker-bogie suspension rover for climbing steps

Abstract

We address the problem of improving mobility of rovers with rocker-bogie suspension. Friction and torque requirements for climbing a single step were considered as performance parameters. The main contribution of the paper is an improved formulation for rover optimization using smooth functions, which enables use of powerful gradient based nonlinear programming (NLP) solvers for finding solutions. Our formulation does not have certain shortcomings present in some earlier formulations. We first formulate the problem of determining optimal torques to be applied to the wheels to minimize (a) friction requirement, and (b) torque requirement, and obtain demonstrably optimal solutions. We then formulate the problem of optimal design of the rover itself. Our solution for climbing a step of height two times the wheel radius is 13% better than that of the nominal rover. This solution is verified to be a local minimum by checking Karush–Kuhn–Tucker conditions. Optimal solutions were obtained for both forward and backward climbing. We show that some earlier formulations cannot obtain optimal solutions in certain situations. We also obtained optimal design for climbing steps of three different heights, with a friction requirement which is 15% lower than that of the nominal rover.