RAIM algorithm for multiple gross errors detection based on Mean Shift model

There is no a good balance between the performance of the detection and the recognition and the calculation efficiency for multiple gross errors in the current receiver autonomous integrity monitoring (RAIM) algorithm. In this paper, the Mean Shift (MS) model is introduced to resolve these problems of the RAIM algorithm. Firstly, the QR parity check method is used to build a sample dataset and the QR calibration vector. Then, the density center is estimated by using the MS model, which is regarded as the MS calibration vector. The distance between the observation vector and the MS calibration vector can be applied for evaluating the reliability of global navigation satellite system (GNSS) observations, and determining the abnormal satellites. Finally, we use the weight coefficient function with a qualitative distance which is derived from the combination of the observation vector, the MS calibration vector and the QR calibration vector to select the abnormal observations and to furtherly promote the performance of detection and recognition of the calculation efficiency of multiple gross errors. The experimental results demonstrate that the gross error discrimination method based on the MS calibration vector has a more sensitive recognition ability in the presence of multiple gross errors. In addition, the new RAIM algorithm can not only obtain the better performance of detection and recognition and the calculation efficiency of multiple gross errors, but also can effectively improve the reliability of single point positioning with multi-system fusion, compared with the least square residual method. © 2022, Editorial Office of Systems Engineering and Electronics. All right reserved.