Integrating directional constraints in motion planning using nonlinear, anisotropic, harmonic potential fields

Abstract

This paper addresses an important problem that has remained, to the best of the author's knowledge, neglected in the motion planning literature. That is the incorporation of directional constraints along with obstacle avoidance in steering a robot to a target zone. The available motion planners deal exclusively with isotropic workspaces that place no restrictions on the direction of motion in their interior. Unfortunately, realistic workspaces do not support the above behavior. Instead, they impose a priori known restrictions on the direction of motion inside a workspace (e.g., ENTER and EXIT signs, direction of traffic along a one-way street). In this paper harmonic field motion planners which can only deal with isotropic workspaces are generalized to incorporate anisotropic workspaces with preimposed directional constraints. Theoretical developments along with simulation results are provided.