Task and Motion Planning for Manipulators Using Linear Temporal Logics

This paper deals with the problem of combining task and motion planning for a manipulator system. We propose a linear temporal logic (LTL) approach which follows the traditional hierarchical planning process, including task planning and motion planning. Firstly, the feasible network topology graph is obtained from the high-level task planning module, and the planning algorithm of search is used to obtain the task sequence. Secondly, the coordination module receives the task sequence to extract the manipulator’s action sequence to make planning suggestions. Furthermore, a lower level planning module is used to receive the suggestions and motion planning is performed using sampled planning algorithms to generate and output the motion trajectory. Finally, we verified the feasibility of our method through simulations and physical experiments. Our method has smoother trajectories and is about 10% more efficient than the traditional method of setting way points.