移動ロボットのモデルベースト故障診断に関する研究-内界センサ系の故障診断機能を有する自己位置推定システム-

摘要

This paper presents a robust self-localization system for mobile robot equipped with the laser/reflectorbased bearing-only sensor (external sensor); the localization system is integrated with a fault detection andidentification (FDI) system of internal sensors, and then it can find the robot position accurately even though faults ofinternal sensors occur. The three failure modes (hard, noise and scale failure modes) of internal sensors are handled inthis paper. The FDI is achieved via interacting-multiple-model method and Kalman filter, while the self-localizationis done via chi-squared hypothesis testing and Kalman filter. Experimental results of a skid-steered mobile robotvalidate the effectiveness of the proposed method.