潜艇适航性监测的卡尔曼滤波器异常数据修正方法

A Kalman filter bad data correction method for monitoring submarine′s seaworthiness

结合潜艇适航性变量监测问题,研究了卡尔曼滤波器的原理和异常行为检测,分析了异常数据对卡尔曼滤波器影响,并通过新息分析和过程信号引入不确定性,给出了一种对异常数据进行剔除并对滤波器进行修正的方法。对潜艇适航性中的横摇进行了数值仿真,给出了经典卡尔曼滤波与改进卡尔曼滤波的对比仿真结果。仿真结果表明:采用改进算法处理,可以有效剔除偏差较大的异常数据;在滤波器已受到异常数据干扰的情况下,改进后的滤波器能够在较短的时间内修正异常数据并重新收敛。

In order to monitor the variations of the submarinels seaworthiness, a Kalman filter bad da- ta correction method was presented. Based on the theories of Kalman filter and the detection of anomalous behavior, the influence of bad data on Kalman filter was detected, and the filter was recovered by introducing some uncertainties to the process noise. The numeric simulation of submarine seaworthiness provided the results of contrast between the classic Kalman filter and the improved Kalman filter. These results demonstrate that the failures caused by bad data can be avoided and overcome.