论文部分内容阅读
UKF作为一种新的非线性滤波方法已在目标跟踪问题中得到应用,在状态的时间更新阶段直接使用非线性模型,不引入线性化误差,而且不必计算Jacobians矩阵.提出了一种基于方根分解形式的带有衰减因子的UKF算法(SRDMA-UKF),算法的方根形式增加了数字稳定性和状态协方差的半正定性.通过衰减因子的引入加强对当前测量数据的利用,减小历史数据对滤波的影响.仿真实验结果表明,该算法与UKF算法相比具有更好的滤波性能.
As a new non-linear filtering method, UKF has been applied in the target tracking problem. The nonlinear model is directly used in the time updating stage of the state without introducing linearization error, and it is not necessary to calculate the Jacobians matrix. The decomposed version of the UKF algorithm (SRDMA-UKF) with attenuation factor, the square root of the algorithm increases the semi-positivity of the numerical stability and the state covariance. The introduction of the attenuation factor enhances the use of the current measurement data, reducing The influence of historical data on filtering The simulation results show that this algorithm has better filtering performance than the UKF algorithm.