Improving the performance of GNSS precise point positioning by developed robust adaptive Kalman filter

Abstract

Global Navigation Satellite Systems (GNSS) Precise Point Positioning (PPP) is a great precision positioning method based on GNSS. PPP based on multi-constellation GNSS uses the extended Kalman filter (EKF). Inappropriately, the measurement outliers and the system's dynamic model might produce mistakes in the positioning accuracy acquired using this method. An adaptive robust Kalman filter (RKF) was recently developed in order to alleviate these errors. In addition, the preceding variance is sometimes used to determine the weights of different categories of observations, which has an effect on PPP performance. A novel developed robust adaptive filtering (DRKF) has been developed to provide better placement results. For the equivalent weight matrix to be generated, an equivalent weight model that is statistically robust by the chi-square test for classification is constructed in this technique. According to prior Kalman filter models employing International GNSS Service (IGS) final clock and orbit products, the performance of GNSS PPP using the DRKF is confirmed in contrast to those produced using the current model. DRKF surpasses traditional robust adaptive filters in terms of positioning accuracy, convergence time, and PPP sturdiness according to the results of this study. In comparison to RKF, DRKF improves the positioning solution by 201 mm with 119.10 %, 77 mm with 164.86 %, and 160 mm with 211.74 % in the east, north, and up (ENZ) directions, respectively. It is recommended to use the newly developed robust adaptive Kalman filter in kinematic mode, especially in urban areas, due to the noticeable improvement in position accuracy.