Integrating Location-Routing with Task and Motion Planning
Yazz Warsame · 2020
This paper presents an approach that combines both location-routing and motion planning. The proposed architecture consists of three phases, namely sampling-based motion planning, facility location and vehicle routing. The sampling-based motion planner uses a probabilistic roadmap to generate a distance matrix. The facility location phase consists of a two-stage clustering algorithm, where the first stage cluster the goals with respect to the vehicles' maximum capacity and the second stage decides the facility location. Finally, in the vehicle routing phase, we use an off-the-shelf PDDL 2.1 solver for the routing problem. Experiments using a second-order vehicle model in obstacle-rich environments showcase the efficiency of our approach.