Research On Fault Diagnosis For Miniature Unmanned Helicopter Navigation System Sensors

Miniature unmanned helicopter(MUH) is widely applied in military and civilian field, and has been one of the most interesting researches in the area of aviation all over the world, with the advantages of small size, low cost, vertical take off and landing ability, hovering ability, low-level flight ability and so on. As the basal parts in MUH and major components in the flight and control system, the performance of the navigation system sensors will impact on flight safety directly.The main research objects are the MUH navigation system sensors in this dissertation, due to low cost, poor stability, and the harsh working environment of high temperature and strong vibration, the sensors are one kind of parts with the lowest reliability in the MUH control system. Aiming at the problem, a method based on phase difference and wavelet packet analysis is used to diagnose the MUH navigation system sensor faults, which adopts the least squares support vector machine(LS-SVM) to classify the fault patterns of sensors. The proposed method is effective and meaningful for the MUH flight experiments. The main work of this dissertation is as follow:1. The current research progress of MUH and the fault diagnosis for MUH sensor all over the world are introduced. And a brief introduction of the navigation system sensors and the fusion algorithm based on complementary filters which is adopted for the flight and control system are given. Finally the content and organization of this dissertation are shown.2. A method of fault diagnosis based on phase difference is proposed, and the model of fault diagnosis based on phase difference is established based on the signal characteristic of MUH navigation system sensors. In the model, the wavelet threshold denoising method is adopted to process the sampled signals, and the estimated phase difference is used to detect faults with correlation analysis. The simulation result shows the method can quickly detect sensor faults.3. Aiming at the defect that the method based on phase difference can not effectively isolate the fault sensor, a method of MUH navigation system sensor fault diagnosis based on phase difference and wavelet packet analysis is proposed. The model based on phase difference is used to detect the sensor faults, after three-level decomposition of original signals associated with fault sensors through wavelet packet analysis, and the fault sensor is isolated with extracted characteristic vectors. Finally the effectiveness of the proposed method is demonstrated by the comparison of with the method based on SVM and wavelet transform and the simulation result.4. The MUH navigation system sensor fault classifier is established based on LS-SVM. For the identification of the fault type of sensor, the characteristic vectors of inertial sensor output signal are extracted through wavelet packet analysis, and used as learning samples to train the constructed classifier so as to realize the mapping relationship between the fault pattern and the characteristic vectors, the fault type of sensor is identified.