Combined Sampling and Optimization Based Planning for Legged Wheeled Robots
Abstract Planning for leggedwheeled machines is typically done using trajectory optimization because of many degrees of freedom, thus rendering leggedwheeled planners prone to falling prey to bad local minima. We present a combined sampling and optimizationbased planning approach that can cope with challenging terrain. The samplingbased stage computes wholebody configurations and contact schedule, which speeds up the optimization convergence. The optimizationbased stage ensures that all the system constraints, such as nonholonomic rolling constraints, are satisfied. The evaluations show the importance of good initial guesses for optimization. Furthermore, they suggest that terrain, collision (avoidance) constraints are more challenging than the robot models constraints. Lastly, we extend the optimization to handle general terrain representations in the form of elevation maps. In IEEE International Confe
|
|