Robust Factor Graph Optimization Integrated Navigation Based on Improved Chi-Square Test

Global Navigation Satellite System (GNSS) outliers and time-varying noise in complex scenarios seriously affect the accuracy and reliability of state estimation. To address this problem, a robust factor graph optimization integrated navigation algorithm based on improved chi-square test is proposed. The binary results of the chi-square test are improved by using the down-weighting strategy of the M-estimator to enhance its stability and eliminate false alarms. The innovation weight is used to sequentially construct a scaling factor to correct the measurement covariance, and outliers are detected based on the new covariance. A robust factor graph based on improved chi-square test is derived and constructed. Integrated navigation simulation and field tests indicate that the proposed algorithm exhibits excellent robustness and positioning accuracy in GNSS degraded environment.