论文部分内容阅读
针对混合估计自适应滤波器的工程应用问题，首先证明了交互式多模型算法（ＩＭＭ） 在一定条件下，其模型输入交互方差可与状态解耦，并给出了两模型下的ＩＭＭ 的模型输入交互方差之间部分解耦及完全解耦的条件，从而将常增益滤波与ＩＭＭ 相结合，提出两模型常增益ＩＭＭ 自适应滤波算法．仿真表明在精度与ＩＭＭ 相当的情况下，计算量减少了约５０％ ，并消除了单模型常增益滤波的有偏性．
In order to solve the problem of engineering application of the hybrid adaptive filter, we first prove that under certain conditions, the interactive multi-model algorithm (IMM) decouples the variance of the model input from the state and gives the IMM model under two models Input partial variance of interaction between the partial decoupling and complete decoupling conditions, which will be combined with the normal gain filtering IMM proposed two models of constant gain IMM adaptive filtering algorithm. Simulation shows that the computational complexity is reduced by about 50% with the same accuracy as the IMM, and the biased filtering of the single-model constant gain filter is eliminated.