基于遗传算法的自动驾驶仪测试性分析

Analysis of Autopilot Testability Based on Genetic Algorithm

摘要:

为了提高导弹自动驾驶仪故障测试与诊断的快速性与准确性,提出了一种基于遗传算法的自动驾驶仪测试性分析方法；此方法通过分析和解决遗传算法应用于故障测试优选时的难点,以测试时间最短为优化目标,得到了故障测试策略的优化方法,在满足测试覆盖性要求的同时,能达到最佳的优化效果；在获得最优测试策略后,利用故障测试获得的知识,简化故障与测试间的相关性矩阵,加快了故障诊断的速度；最终获得了故障测试与诊断一体化的方法；实验结果表明,该方法能有效提高故障测试与诊断的速度,并且易于实现。

Abstract:

In order to improve the efficiency and accuracy of fault test and diagnosis on missile autopilot, a method based on the genetic algorithm was put forward to analyze the testability of autopilot. Through analyzing and solving the the problem on how to make genetic algorithm apply to the optimization of test items, and taking the shortest test time as the optimization goal, this method got the best fault test strategy. The strategy could achieve the best optimization and meet the requirement of test coverage. Then the knowledge got from fault test was used to simplify the dependency matrix between fault and test, and the diagnosis speed was increased. Finally a method that combines fault test and fault diagnosis was got. The experiment result shows that the method can increase the speed of fault test and diagnosis. At the same time, the method is easy to implement.