论文部分内容阅读
推广了估计时不变参数的单新息修正技术，提出了多新息辨识方法．该方法可以抑制坏数据对参数估计的影响，具有较强的鲁棒性．分析表明多新息方法可以跟踪时变参数，计算量也较遗忘因子最小二乘法和卡尔曼（Ｋａｌｍａｎ）滤波算法要小．仿真结果说明多新息算法估计系统参数是有效的．
The single-rate-of-interest correction technique is introduced to estimate the invariant parameters of the estimation, and a multi-innovation identification method is proposed. This method can restrain the influence of bad data on parameter estimation and has strong robustness. The analysis shows that the multi-innovation method can track the time-varying parameters, and the computational cost is also smaller than the forgetting factor least square method and the Kalman filtering algorithm. Simulation results show that multi-innovation algorithm to estimate system parameters is valid.