Fault diagnosis of certain missile based on dynamic test

ATS(automatic test system) is one of the major martial maintenance equipments for certain missile. But the available missile ATS prolongs with traditional static test method basically, and exists some shortcomings. This paper proposed fault diagnosis based on dynamic test to upgrade certain missile ATS. Certain missile´s dynamic test is realized by simulating the missile´s flight course. According to the missile´s characters displayed at each stage of boost, cruise, fall, swoop in real flight, we issue control commands to course channel, pitching channel and incline channel at the right moment respectively. At the same time, simulate the missile´s pose change by controlling the turntable´s moving. So we can simulate the missile´s flight and attack course factually. A novel residual error energy characteristic vector method based on WPT (wavelet packet transform) is proposed, and is used to extract fault feature. Then import the fault residual error energy characteristic vector to the BP neural network to diagnosis faults. Experiments prove that this method is effective.