自适应渐消有偏扩展卡尔曼滤波在目标跟踪中的应用

针对机动目标跟踪过程观测矩阵病态导致扩展卡尔曼滤波算法跟踪效果不佳的问题,提出一种自适应渐消有偏扩展卡尔曼滤波算法。该算法以扩展卡尔曼滤波为基本框架,并借鉴Gauss-Markov模型的思想以解决观测矩阵病态问题。算法根据状态估计均方误差最小条件求得有偏因子,以降低病态观测矩阵对滤波估计的影响;根据滤波发散判据提出一种新的渐消因子估计方法,以实时调整预测协方差矩阵,从而改善滤波增益并有效提高目标跟踪精度。仿真结果表明,改进算法比传统扩展卡尔曼滤波对目标跟踪的精度有较大提高,同时稳定性更好。

Application of the Adaptive Fading Biased Extended Kalman Filter in Target Tracking

An adaptive fading and biased extended Kalman Filter(EKF)algorithm is proposed for the lower tracking performance which is caused by the ill-conditioned measurement matrix under the target maneuvering conditions. The algorithm uses the EKF framework and draw on the experience of Gauss-Markov model to solve the ill-conditioned problem of the observation matrix. For the algorithm,a bias factor is obtained by minimizing the mean square error of state estimation to reduce the influence of the observation matrix for the filtering estimation. Furthermore,a new fading factor estimation method is proposed by the filter divergence criterion to adjust the prediction covariance matrix in real time,then to improve the filter gain and the tracking accuracy. The simulation results show that the improved algorithm is with higher tracking accuracy than that of the traditional EKF for target tracking. Also,the tracking stability of the proposed algorithm outperforms the traditional one.