Fault-Tolerance of Linear Quadratic Optimal Systems

Abstract

This paper investigates fault-tolerance of linear quadratic optimal systems. Fault-tolerance is defined as the bound of allowable sensor and actuator gain degradations to maintain stability. Quantitative results for fault-tolerance are obtained using Lyapunov matrix equation solutions. A relation is established between the fualt-tolerance and the prespecified stability degree ¿ and weighting matrix Q in cost function. An iterative algorithm is developed to design the system of the highest fault-tolerance.