基于灰色预测和KID离群点检测的AUV故障诊断

Outlier Detection Algorithm Based on Gray Prediction and KID for AUV Fault DiagnosisZhou Yue1,2,3Wang Haoming2Guo Wei3

摘要:

提高故障诊断能力对于确保水下机器人AUV系统的稳定运行具有重要意义。针对水下机器人推进器系统,提出一种基于离群点检测的AUV故障检测方法。首先,将传感器采集的数据进行灰色预测处理；然后,提出了一种结合K-mean和DBSCAN的改进迭代聚类(Iterative K-mean DBSCAN,IKD)算法进行离群点检测；最后,与K-mean和DBSCAN算法相比,仿真实验结果表明基于灰色预测和KID离群点检测算法的故障检测准确率高,能够有效地实现水下机器人AUV的无监督故障诊断。

Abstract:

It is very important to improve the fault diagnosis ability to ensure the stable operation of the autonomous underwater vehicle (AUV) system. A fault detection method based on detection is proposed in this paper for AUV propeller system. Firstly, the gray prediction algorithm is used to complete data preprocessing for the data collected by sensors. Secondly, an improved iterative clustering algorithm combining K-mean with is put forward to detect outliers. Finally, the simulation results demonstrate the outlier detection algorithm based on gray prediction and KID has higher accuracy, compared with K-mean and DBSCAN algorithms, which could implement a comprehensive unsupervised fault diagnosis of AUV.