一种基于生成对抗网络的卫星异常检测方法

An Anomaly Detection Method Based on Generative Adversarial Network

针对卫星地面测控中心在异常检测时面临的遥测数据不平衡和缺乏异常标签等问题,提出了一种基于时序生成对抗网络的异常检测方法.首先对卫星遥测数据进行预处理,剔除原始数据中的噪声和野值.然后使用长短时记忆网络构建生成模型的生成器和判别器,使得模型可以学习到历史数据的时间依赖关系.采用改进的生成对抗损失函数,使得生成模型在训练时可以保证生成序列与输入序列的潜在空间分布一致.最后,使用残差作为测试序列的异常分数,通过阈值自适应方法判断测试序列是否异常.经真实卫星遥测数据进行实验验证,表明该异常检测方法具有较好的有效性.

For satellite telemetry data, imbalance of distribution and lack of anomaly tags are main problems. Therefore, usual anomaly detection method is not effective in this respect. An anomaly detection method based on timing generation countermeasure network is proposed. Preprocessed to eliminate the noise and outliers, the satellite telemetry data is used to train the generator and discriminator, so that the time dependence of historical data can be learned. Finally, the adaptive threshold method is used to determine whether the residual of generator is abnormal. The experimental results show that the anomaly detection method is effective.