基于CS-Jounce模型的船摇滤波与预报方法

根据航天测量船船摇数据特性,提出了基于“当前”统计Jounce模型(“current” statistical Jounce model, CS Jounce)和最大相关熵卡尔曼滤波器的滤波及预报方法。在分析船摇数据周期特性的基础上,使用频域滤波方法分析了船摇数据噪声,得出其非高斯白噪声,指出去船摇数据四阶差分量周期性显著减弱。为克服船摇动力学建模复杂的问题,引入机动目标跟踪理论,针对目前模型阶数不够高的问题,结合“当前”统计的思想,推导了CS-Jounce模型。针对船摇数据噪声非高斯白噪声的问题,引入了最大相关熵卡尔曼滤波器。实测数据验证表明,该方法能很好地跟踪船摇数据,滤波精度较高,三步预报均方根误差小于5″。同时,该方法能较好地估计和预报船摇角度及角加速度数据。

Filtering and prediction method based on CS-Jounce model for ship swaying data

According to the characteristics of swaying data of spacecraft tracking, telemetering and control ships, a filtering and prediction method, based on a “current” statistical Jounce model model, and the maximum correntropy Kalman filter is proposed. Based on the analysis of the periodic characteristics of ship swaying data, the data noise is analyzed by using frequency domain filtering methods, and it is a non-Gaussian white noise. It is pointed out that the fourth order difference of the ship swaying data is significantly weakened periodically. To tackle the complication of ship rocking dynamics modeling, the maneuvering target tracking theory is introduced. In view of the problem that the current model order is not high enough, the CS-Jounce model is derived based on the “current” statistics. Aiming at the problem of non Gaussian white noise of ship swaying data, the maximum correntropy Kalman filter is introduced. The experimental results based on actual data show that the proposed method can track the ship swaying data well. The filtering accuracy is high and the mean square error of the three step prediction is less than 5 seconds of arc. Meanwhile, the proposed method can accurately estimate and predict the ship's angular velocity and angular acceleration.