Analysis and Comparion of Robust Least Squares Estimation Based on Multi-constellation Integrated Navigation

To solve the problem of the least squares estimation does not have the ability to resist the gross errors in global navigation satellite system (GNSS), the theory of robust least squares estimation is introduced, three methods (Huber, IGG, IGGIII) to construct the equivalent weight are expounded and the determination of initial parameter of the equivalent weight is discussed. The results of simulation based on GPS/BDS multi-constellation show that the robust least squares estimation can ensure the accuracy of navigation and positioning effectively, the estimator with IGGIII equivalent weight method presents optimal solutions, furthermore the performance of the robust estimation method based on median parameter estimate is better than the individual robust estimation.