Support Vector Machine Observer Based Fault Detection for Reconfigurable Manipulators

A novel fault detection scheme for reconfigurable manipulators actuators based on support vector machine(SVM) observer which has good ability to estimate any functions is proposed to estimate the unknown term and uncertainty term. The reconfigurable manipulators system is fault-free when the error between true value and predict value is less than the given threshold, otherwise, the system is in failure. Then the fuzzy logic systems are proposed for approximating the interconnection term by using adaptive algorithm. Based on Lyapunov stability theorem, the system stability can be verified. And the simulation results illustrate the effectiveness of the proposed fault detection schemes.