Research Of Condition Monitoring For Sensors Of Autonomous Underwater Vehicle

The research of the autonomous underwater vehicle is one of the most important parts of the naval national defense and ocean development stratagem. To guarantee the higher efficiency and reliability in the research, development, design and manufacture, and finish the motion control, mission and safely homing in the underwater experiments and actual combat, it is necessary to monitor the condition of some components in the AUV system.In this article, according to the analysis of the sensor system in the AUV's control system, a brief and feasible method of sensor grouping is presented, and a structure of condition monitoring system based on the intelligent method is established whose kernel is diagnoses and signal resume model. The method mainly constructs the diagnoses and signal resume models of each sensor of the AUV. By monitoring the sensor signal under the patter of the cruising and hunting, it diagnoses whether some faults have occurred. If some sensors have faults, it will resume the signal to the fault sensor in order to ensure the navigation security.In the constructing of the monitoring module, the technology of the combination of the data fusion and neural network is adopted which based on the RBF, and the method of confirming the center and radius is using. The technology also adjusts parameters to perfect the network structure and performance based on the grads method. By combining the movement data which the controller output and the diagnosing result which the diagnosing model output, the technology resume the signal. Thus offers the proof to complete mission and return safely. The result of the computer simulation shows that the method is feasible and the condition monitoring system is effective.