基于Elman网络的传递对准容错联合滤波器设计与仿真

Transfer Alignment Using Fault-tolerant Federated STF Based on Elman Network

针对用于快速传递对准的Kalman滤波器阶数高,计算量大,滤波更新率低,鲁棒性差及对准精度不高等问题,采用联合强跟踪Kalman滤波器进行快速传递对准。提出一种基于模糊加权系数的误差方差阵估计方法,以提高传统强跟踪Kalman滤波算法的精度。在此基础上,设计了联合强跟踪Kalman滤波器的结构和算法。基于提高无故障子滤波器的鲁棒性来提高联合滤波器的快速重构能力考虑,同时兼顾子滤波器的精度和计算稳定性,提出利用改进的Elman网络进行信息分配系数的自适应调节,以实现融合信息在各子系统中的自适应分配。仿真结果表明,该滤波器不仅提高了解算速度,而且提高了系统对准精度、故障鲁棒性和快速重构能力。

Considering the huge calculation burden on the computer brought by the high dimension number of the centralized Kalman filter for the rapid transfer alignment and the poor robustness of Kalman filter,the federated strong tracking Kalman filter is presented in this paper. Firstly,an estimation method of variance matrix based on fuzzy weighting coefficient is proposed to improve the accuracy of the traditional strong tracking filter. Then the structure and algorithm of federated strong tracking Kalman filter are designed. To improve the reconfiguration ability of federated filter by improving the robustness of local filters,and to improve the accuracy and calculation stability of local filters,an improved Elman network is proposed to adjust information sharing coefficients adaptively according to some experiential rules. Simulation results show that this filter not only improves the calculation speed,but also improves alignment accuracy,the system robustness and rapid reconfiguration ability.