一种具有SA预报能力的GPS KALMAN滤波器

A KIND OF GPS KALMAN FILTER WITH SA PREDICTION

对GPS定位中的SA误差作了系统辨识。表明：SA误差可用时变AR（p）模型表述，AR（p）的时变参数可由RLS算法结合F判据建立的AR（p）模型辨识机获得。用该方法获得的模型对SA误差进行预报。在剔除野值后精度在10m之内。还提出了一种将以AR（p）模型表述的SA误差加入GPSKALMAN滤波器的方法，可得到一种具有SA预报能力的GPSKALMAN滤波器。

This paper first establishes a model of GPS SA error with system identification. The result shows that the AR(p) and that model of GPS SA is time-varied.The time-varied parameters of the model can be estimated by a RLS and F judgment estimator. The accuracy of prediction with this model is less than 10m after preprocessing. Then, by integrating this into the GPS KALMAN filter we obtain a kind of GPS KALMAN FILTER with SA prediction.