Fault diagnosis for missile autopilot based on GSA-SVM

Due to the blind selection for the kernel function parameters and penalty factor parameters by Support Vector Machine (SVM), a method based on gravitational search algorithm (GSA) to optimize the SVM parameters was proposed and was used for the fault diagnosis of analog circuit in a missile autopilot. This kind of method can avoid falling into the local optimal problem during the process of optimizing the SVM parameters and then can achieve the global optimal solution. The experimental results showed that the SVM optimized by this method had a higher diagnosis precision and robustness than that optimized by particle swarm optimization (PAO) and genetic algorithm (GA). Moreover, this method proposed in the paper can complete the fault diagnosis well.