Combining harmonic functions and random sampling in robot motion planning: a lazy approach

This paper presents a lazy procedure that enhances the performance of a robot motion planning method, called PHM, that uses a potential-field approach based on harmonic functions together with a random sampling scheme. The harmonic functions used to guide the solution path are computed over a 2^d-tree decomposition of a dimensional configuration space that is obtained with probabilistic cell sampling. This paper proposes a lazy variant of the PHM planner that eliminates, reduces or delays as much as possible any time-consuming computation. The proposed approach , therefore, makes the planner computationally more efficient