论文部分内容阅读
集中式融合滤波对于噪声相关非线性多传感器系统很重要。首先,在扩展Kalman滤波器(EKF)的基础上,利用矩阵求逆引理推导出噪声相关的EKF的一种信息滤波器形式;然后,根据矩阵相似变换理论将其等价分解为具有局部通信的微观滤波器形式。与现有的集中式融合算法相比,新方法保持了相同融合精度的同时,还具备了部分信息滤波器的优良数值计算特点。最后,通过理论分析和计算机仿真相结合的方法来验证了新算法的有效性。
Centralized fusion filtering is important for noise-dependent nonlinear multisensor systems. First, based on extended Kalman filter (EKF), an information filter of noise-related EKF is deduced by matrix inversion lemma. Then, according to the matrix similarity transformation theory, it is equivalently decomposed into two parts: local communication Micro-filter form. Compared with the existing centralized fusion algorithm, the new method retains the same fusion precision, and also possesses the excellent numerical calculation of some information filters. Finally, the effectiveness of the new algorithm is verified by a combination of theoretical analysis and computer simulation.