A Multi-layered Synergistic Approach to Motion Planning with Complex Goals

This paper describes an approach for solving motion planning problems for mobile robots involving temporal goals. The temporal goals are described over subsets of the workspace (called propositions) using temporal logic. The approach uses an instantiation of a multi-layered synergistic planning framework that has been proposed recently. In this framework, a high-level planner constructs high-level plans using a discrete abstraction of the system, the temporal logic specifications and the low- level exploration information. A low-level sampling-based planner uses the suggested high-level plans and the dynamics of the system to explore the state-space of the system for feasible trajectories satisfying the specification. The construction and exploration of the discrete abstraction are critical issues that affect the overall performance of the approach. A geometry-based approach for constructing the abstraction, and a lazy high-level search technique for its exploration are discussed. The proposed techniques result in computational speedups of close to 10 times over earlier approaches for second-order nonlinear robot models in challenging workspace environments with obstacles and for a variety of temporal logic specifications.