Algebraic Fault Detection and Identification for Rigid Robots

This paper presents a method for algebraic fault detection and identification of nonlinear mechanical systems, describing rigid robots, by using an approximation with orthonormal Jacobi polynomials. An explicit expression is derived for the fault from the equation of motion, which is decoupled from disturbances and only depends on measurable signals and their time derivatives. Fault detection and identification is then achieved by polynomial approximation of the determined fault term. The results are illustrated for a faulty SCARA.