专家系统与测试技术相结合的诊断维护系统

本文介绍的诊断维护系统ＩＴＭＤＭＳ将专家系统技术与传统的测试知识相结合。该系统包括两类知识：浅层知识和深层知识，知识用一阶谓词表示。推理过程分三步完成：（１）浅层推理；（２）深层推理；（３）浅层推理。知识维护手段使系统的知识在使用中不断完善。该系统的框架可适用于任意复杂的数字系统的诊断维护。

Robust Diagnosis for Sensor Failure

An approach for sensor fault diagnosis in stochasitc control system is proposed. Fault detection is realized on the basis of Kalman filter innovation. The location of sensor fault is obtained by utilizing the decision function of error estimation. Simulation results indicate that the approach is feasible and efficient.