設備自動点検システム構築に向けた異種ロボット群のタスク割り当て効率化

抄録

Currently, instrumental inspections in factories are carried out manually, and it is strongly required to automate the inspections and/or improve the task efficiency. In view of the needs, we consider we address optimal task assignment with a heterogeneous robotic group. First, constraints that have to be met in the robotic inspection task are enumerated and they are represented as logical constraints. Then, the logical constraints are transformed into linear inequalities for appropriately defined binary variables, and the task assignment problem is reduced to an integer programming problem. We further propose a method to reduce the computational burden by dividing the workspace and splitting the optimization problem into an upper layer that allocates regions and a lower layer that allocates targets and tasks. To guarantee the feasibility of the lower layer optimization, new logical constraints are introduced in the upper layer. Finally, a numerical example simulating a real-world working environment confirms the effectiveness of the solution and the reduction in computation time due to the domain partitioning.