【摘要】 针对多尺度带乘性噪声系统,在多尺度最优滤波融合的基础上,进行状态最优固定域平滑算法的研究。通过推广得到的平滑算法需要大量的局部传感器参数,而分布式多尺度滤波融合后不能保留这些信息。针对这一弊端对算法进行改进,推导出仅使用融合后的一步预测及滤波值的平滑算法。该算法在线性最小方差意义下是最优的。计算机仿真验证了算法的可行性。更多还原

【Abstract】 Based on multiscale optimal filtering fusion,an optimal state fixed-interval smoothing algorithm is developed for systems with multiplicative noise.The smoothing algorithm obtained by generalization requires a great many local sensors’ parameters while the information cannot be reserved through distributed multiscale filtering fusion.In order to solve this defect,this paper proposes a smoothing algorithmonly using one-step prediction and filtering value.The algorithm is optimalin the sense of linear minimum-variance.The feasibility of the algorithmis shown by computer simulations.更多还原