信頼性理論を用いたロボットの作業環境整備法と動作生成法

Environmental support is one of the effective approaches for robots to do tasks efficiently in office and home environments. In order to accomplish robust and rapid task realization of robots, both the environmental support and robot motion should be designed carefully. Authors have proposed the integrated methodology of designing two factors by using reliability theory. In this paper, we realize environmental support by attaching "artificial marks with memory" to all the objects in the working environments. We take the following steps: (a) preparing four motion primitives for robot motion generation. (b) Classifying "task realization method" into four from the viewpoint of environmental support and robot motion generation. (c) Programming motion sequence for each class by combining the four primitives in (a), after getting task specification. (d) Evaluating four classes by using reliability theory. We show the effectiveness of the proposed methodology through realizing a sliding door-closing task and a faucet-closing task(see video) with a real robot system.