Optimal batch recursive data compression

A recursive linear optimal estimation method is presented which processes a batch of measured data at each estimation point. Due to this feature the new method serves as a data compression scheme. This scheme is an integrated one in the sense that the optimization is performed on a prefilter/estimator combination. The method is derived and compared for accuracy performance and complexity with the Kalman filter. It is found that a considerable saving in computational effort, with a minimal degradation of accuracy, is achieved using this method. An example is presented and efficient algorithms to implement this method are developed.