卡尔曼滤波与多传感器数据融合研究

Research on Kalman Filter and Multisensor Data Fusion

为了解决多传感器数据融合对目标状态进行估计时会出现噪声影响这一问题,采用一种在机动目标跟踪中性能较好的卡尔曼滤波算法(Kalman filtering,KF)。详细介绍了离散时间系统的数学模型,进而推导出卡尔曼时间更新方程和卡尔曼量测更新方程,以及测量变量与状态变量间的关系矩阵的传递过程,各个系数矩阵的重要意义。最后,基于一个航迹融合模型,建立其系统状态空间模型的矩阵方程,通过仿真研究其滤波前后目标的轨迹与误差变化,进而验证了该算法的有效性。

To solve problem of noise influence during multi-sensor data fusion estimating target state, Kalman filtering algorithm(KF) with good performance in maneuvering target tracking can be adopted. The paper introduces mathematical model of discrete-time system in detail, then deduces Kalman time renewal equation and Kalman measurement renewal equation, and transfer process of relationship matrix between measurement variables and state variables, and significance of each coefficient matrix. Finally, based on track fusion model, it establishes matrix equation of its system state space model, and studies track and error changes of target before and after filtering by simulation, and verifies effectiveness of the algorithm.