基于滑模观测器的机器人系统传感器故障诊断

SLIDING MODE OBSERVERBASED FAULT DIAGNOSISFOR ROBOT SENSORS

本文基于变结构滑模状态观测器，提出了一种在模型含有不确定性误差的情况下，机器人控制系统传感器故障的鲁棒诊断方法．通过滑模观测器与传感器的输出构造残差信号，然后从残差中提取故障信息，实现故障的检测与分离．实验结果验证了该方法的有效性．

In this paper a nonlinear sliding mode observerbased fault detection and isolation scheme for robot sensors is presented. The observer generates the residuals which are robust to model uncertainty and sensitive to specific faults, and fault detection and isolation can be achieved by the characteristics of the residuals. Experimental results show the feasibility of the proposed approach.