针对可重构机械臂系统位置传感器和速度传感器多故障,提出一种主动取代分散容错控制方法.基于可重构机械臂的模块化属性,设计正常工作模式下的分散神经网络控制器.利用微分同胚原理将子系统结构进行非线性变换,将传感器故障转化成伪执行器故障,设计分散滑模观测器以对多传感器故障进行实时检测,并利用其输出信号取代故障传感器信号,实现了多传感器故障情形下可重构机械臂的主动容错控制.仿真结果表明了所设计的容错控制方法的有效性. Aiming at the multi-fault of reconfigurable robotic system position sensor and speed sensor, a decentralized distributed fault tolerant control method is proposed. Based on the modularized property of reconfigurable manipulator, decentralized neural network controller is designed in normal working mode. Homomorphism transform subsystem structure nonlinearly, transform sensor fault into pseudo-actuator fault, design sliding-mode scattered observer to detect multi-sensor fault in real time and replace fault sensor signal with its output signal Active Fault-Tolerant Control of Reconfigurable Robotic Arm in Sensor Fault Condition Simulation results show the effectiveness of the proposed fault-tolerant control method.