汎用ロボットと環境脱着ロボットアーム群によるタスク委任スケジューリング

抄録

This study aims to solve the sequential execution bottleneck caused by a single general-purpose robot performing multiple tasks. In our approach, the robot achieves parallelization by delegating tasks to detachable arms, which can be placed in and retrieved from the environment as needed. Considering the overhead of arm placement and retrieval, we define the “Task Delegation Scheduling Problem”, where the system selects whether the robot executes a task itself or delegates it to an arm. Through simulations, we present guidelines for system design and task operation and analyze the mechanisms of optimal strategies based on task structure. Furthermore, we conduct a proof-of-concept with a real robot and discuss the application scope of this method.