A Study on Moving Window Adaptively Weighting Estimation Method

The paper gives a new moving window random adaptively weighting adaptively estimation method. This new method adopts the concept of weighting estimation with moving window to control predicted residuals and the observational residuals for resisting the disturbances on observations and the system model errors. The theories and methods of moving window adaptively weighting estimation can estimated the covariance matrix of the observational residuals and the innovation vectors. From the experimental results and its comparisons that can demonstrate this proposed moving window random adaptively weighting adaptively estimation method cannot only control the observational error vector and predicted residual vector of the covariance matrix, but also it can effectively resist the disturbances of system error. By the calculation it can be obtained this method is much better than the traditional Kalman filtering method and adaptive filtering method.