Abstract: Given that the precision-guided airborne missile is a sophisticated system, timely and accurate fault identification and diagnosis for missiles are crucial measures to address and prevent risks. Due to the complexity of the missile failure mode, it is difficult to identify and diagnose them. Taking a typical testing problem of a airborne missile as an example, this paper proposed a BP-neural network-based approach of automated fault identification and diagnosis for missiles. By collecting and sorting the missile test data, a data sample was formed. The learning and judgment ability of the neural network system was used to automatically identify and diagnose missile faults. A Matlab neural network toolbox was adopted to verify stimulation for this approach, which showed that the technology could identify and diagnose missile faults quickly and accurately.

Key words: missile testing, BP neural network, automatic diagnosis expert system