A simple recursive algorithm for diagnosis of abrupt changes in signals and systems

Abstract

We address the problem of detecting and isolating abrupt changes in signals and systems with random disturbances. An asymptotic optimal solution to this problem, which has been proposed in previous works, leads to the number of arithmetical operations at time t which grows to infinity with t. Moreover, the proposed criterion of optimality does not impose a constraint on the probability of false isolation when the change time t/sub 0/>1. In this paper we propose another more realistic criterion, establish a new simple recursive change detection/isolation algorithm and investigate its statistical properties. This algorithm can be recommended for on-line implementations.