Fault-Tolerant Control Against Performance Degradation of Actuators for a Robotic System with Guaranteed Prescribed Performance

Haifeng Huang,Wei He,Shuang Zhang,Weicun Zhang
DOI: https://doi.org/10.1109/icarm.2018.8610733
2018-01-01
Abstract:In this paper, the fault-tolerant control problem of an uncertain multi-input and multi-output (MIMO) robotic system is investigated. Neural networks based on Gaussian radial basis functions are used to compensate for the dynamic uncertainties. In the control design, prescribed performance issue is transformed into the time-varying constraints problem via an appropriate way, and then barrier Lyapunov function (BLF) is utilized to rigorously guarantee not only the steady state performance but also the transient performance. Partial loss of effective of the actuators is considered and compensated by the proposed approach. Simulation on a two-link manipulator have confirmed the theoretical results that the presented adaptive neural control scheme can ensure the prescribed tracking performance even when some unknown actuator failures occur to the unknown system.
What problem does this paper attempt to address?