Motion planning of a biped robot on uneven terrain scattered with obstacles by searching optimal location of via points

This article presents a new global path planning method between the initial and goal location for a 14-DOF biped robot on uneven terrain scattered with obstacles. The desired global path is generated by searching the optimal position of the via points on rough terrain. The global pathfinding problem is formulated as an optimization problem by minimizing a path integral over a scalar field called an interpolant uneven terrain energy field (UTEF) function. The global path that is searched is parametrized in terms of via points, and the optimal position of via points is obtained by optimization using a genetic algorithm (GA). The UTEF is defined numerically at discretized terrain nodes using the biped robot's dynamics. Mathematically UTEF incorporates the biped robot's local dynamics at discretized terrain nodes. It is found that the dynamics-based path is more optimal in the sense of path length, stability(passing through safer regions), and closeness to the obstacles. The effectiveness of the proposed global pathfinding method is demonstrated by several numerical experiments in MATLAB, showing the usefulness of the proposed path-planning scheme.