Research On Decentralized Fault Tolerant Control For Reconfigurable Manipulator With Multiple Concurrent Failures

With the science technology rapidly developing, modern system is developing towards the large scale and the complication. It owns the broad application prospect in the mission people can not directly participate in, such as the aerospace and the precise assembly. Among this kind of systems, the application area of reconfigurable manipulator becomes broader and gets more and more investment to improve the requirement of life and reliability. So it is necessary for reconfigurable manipulators system with concurrent actuator and sensor failures to make research on the fault detection and fault-tolerant control methods. For the occurrence of concurrent actuator and sensor failures, this paper presents two strategies, namely, the passive fault tolerant control and active fault tolerant control combining the fifth subtopic "reconfigurable manipulators fault-tolerant control" theoretical research of national natural science foundation of China-Research on Modularized Collaborative Control for Reconfigurable Manipulators with Task Constraint Inspiration. The research content mainly includes several aspects as follows:1. First, this paper introduces the background and signification of my study. Then it presents the brief summary of the current state on the reconfigurable manipulators and the fault detection and fault-tolerant control methods respectively.2. In order to reflect the endemic modular property of reconfigurable manipulators system, through Newton-Euler iteration algorithm and previous experience this paper makes the brief introduction about the establishment of the dynamical model of reconfigurable modular robot. Based on the accumulation of our team’s theoretical knowledge, apply the decentralized control to the control algorithm and the dynamical model of reconfigurable manipulators subsystem can be gained.3. For the occurrence of concurrent actuator and sensor failures (or a single one), the passive fault tolerant control method based on the multi-step time-delay is introduced. Based on the existing achievements, the decentralized backstepping neural network controller for the normal system is applied to the system without the faults. When the faults in the system occur, the multi-step time-delay technique is adopted to compensate the concurrent failures and realize the fault-tolerant.4. Combining the defects existing in the passive fault tolerant control and the shortage existing in the active fault tolerant control of the third chapter, the dynamic output feedback active decentralized fault tolerant controller is presented. The decentralized proportional-integral observer (DPIO) is designed to estimate the faults in the system online. This method owns the characters of not needing fault detection and isolation, compensating the shortage of the normal active fault tolerant controller.Finally, the conclusion for this paper combining my own research findings together with the further study works come to the end of the paper.