Integrating HTN planner in cleaning-security robot: Handling planning with memory and problem template

As the pioneers of robots for family services, the current cleaning and security robots cannot implement high-level and multi-task planning; an efficient solution to adapt the dynamic environments of household work is desired. To deal with this issue and enhance the flexibility of robots with partial observability, a HTN (Hierarchical Task Network) -based semantic planner integrating the robot´s control system is developed in this paper. An approach for handling the planning with incomplete information and exception is proposed by using the robots´ memory and problem template. Finally, several experiments demonstrate that the approach is applicable in an unstructured house environment.