RESEARCH OF THE CONDITION MONITORING SYSTEM FOR SENSORS OF AUV

摘要

A modeling method for diagnosing the faults and restoring the uncertain signals of sensors was proposed, which used a combined RBF network and resolved the problem of multi-sensors coupling of Autonomous Underwater Vehicle (AUV). In the common controller system, each sensor had an RBF identification network for its own, and by comparing the dispersion of actual output and the model output with an experiential dthreshold on a prescribed period of time, it could detect the fault occurring on the sensor being monitored. All sensors were classified according to the signal comparability, so the signal of the fault sensor could be correct by the RBF restoration network, which consisted of the sensors with similar output. The results of the computer simulation by actual experiment data of a certain AUV shows that the combined RBF network using in the multi-sensors fault diagnosis and signal restoration is effective and prove that the condition monitoring model proposed in this article is feasible.