基于最大相关熵的GPS/INS组合导航融合算法研究

Research on GPS/INS Integrated Navigation Fusion Algorithm Based on Maximum Correlation Entropy

目前组合导航系统中常用卡尔曼滤波器进行信息融合,它在一定程度上克服了组合导航系统在实际工作中的不确定情况,如测量值易被无法量测的野值污染的实际问题。为优化传统卡尔曼滤波算法性能,提出基于最大相关熵的卡尔曼滤波(MCCKF)算法。该算法是在最大相关熵准则(MCC)和加权最小二乘(WLS)的思想下进行推导得到的。通过Matlab软件对GPS/INS组合导航融合理论建立数学模型,并对其进行模拟和分析,与传统卡尔曼滤波算法所得出的结果相比,结果表明基于最大相关熵的卡尔曼滤波算法具有更高的精度和更强的鲁棒性。

At present,Kalman filter is commonly used in integrated navigation system for information fusion.It overcomes the uncertainty of integrated navigation system in practical work to a certain extent,such as the actual problem that the measured value is easily polluted by outliers that cannot be measured.To optimize the performance of the traditional Kalman filter algorithm,a maximum correlation entropy-based Kalman filter(MCCKF)algorithm is proposed.The algorithm is derived based on the idea of Maximum Correlation Entropy Criterion(MCC)and Weighted Least Squares(WLS).The mathematical model of GPS/INS integrated navigation fusion theory is established by Matlab software,and it is simulated and analyzed.Compared with the results obtained by the traditional Kalman filter algorithm,the Kalman filter algorithm based on maximum correlation entropy has higher accuracy and more robustness.