Monte-Carlo Tree Search with Prioritized Node Expansion for Multi-Goal Task Planning

Kai Pfeiffer,Leonardo Edgar,Quang-Cuong Pham
2023-07-24
Abstract:Symbolic task planning for robots is computationally challenging due to the combinatorial complexity of the possible action space. This fact is amplified if there are several sub-goals to be achieved due to the increased length of the action sequences. In this work, we propose a multi-goal symbolic task planner for deterministic decision processes based on Monte Carlo Tree Search. We augment the algorithm by prioritized node expansion which prioritizes nodes that already have fulfilled some sub-goals. Due to its linear complexity in the number of sub-goals, our algorithm is able to identify symbolic action sequences of 145 elements to reach the desired goal state with up to 48 sub-goals while the search tree is limited to under 6500 nodes. We use action reduction based on a kinematic reachability criterion to further ease computational complexity. We combine our algorithm with object localization and motion planning and apply it to a real-robot demonstration with two manipulators in an industrial bearing inspection setting.
Robotics
What problem does this paper attempt to address?
The paper aims to address computational challenges in multi-objective task planning, particularly in robotic operations. Specifically, the authors propose a new algorithm based on Monte Carlo Tree Search (MCTS) that optimizes multi-objective symbolic task planning through Prioritized Node Expansion (PNE). This method is particularly suitable for tasks that require handling multiple sub-goals, such as industrial bearing inspection scenarios. Main issues: - How to effectively plan the sequence of actions for a robot in multi-objective tasks? - How to simplify computational complexity in the presence of multiple sub-goals? - How to ensure a feasible solution is found with limited computational resources? The authors address these issues through the following points: 1. **Proposing the PNE algorithm**: This algorithm reduces unnecessary exploration and improves search efficiency by prioritizing the expansion of nodes that have already achieved certain sub-goals. 2. **Linear complexity**: The time complexity of the algorithm is linearly related to the number of sub-goals, allowing it to remain efficient even for larger problem scales. 3. **Action Reduction (AR)**: The algorithm uses kinematic reachability criteria to remove infeasible actions, further reducing the complexity of the search space. 4. **Validation in practical applications**: The effectiveness of the algorithm is demonstrated in a real-world scenario of a dual-arm robot inspecting bearings. In summary, this paper is dedicated to developing a fast and efficient multi-objective task planning method, particularly suitable for complex tasks in industrial automation.