Knowledge based diagnostics for flight control

Describes a knowledge based approach to the health condition monitoring problem of a large complex system. The specific issues concerned include failure detection, isolation and identification of aircraft flight control systems. Few papers have substantially addressed the knowledge based approach for failure detection and diagnosis for aircraft flight control systems. The authors offer a novel approach towards this problem. The approach makes use of a hypotheses generation and test technique to isolate and identify the responsible faulty component or subsystem in a system. Its specific characteristics include, object-oriented hierarchical representation of a system, the concept of 'constraint' which describes the I/O function relationship of a system component and 'constraint suspension'. The entire methodology consists mainly of a low level constraint evaluation process and a high level scheduling rule based reasoning process in which certain heuristic knowledge has been incorporated. The remaining part of the paper details the approach.