Adaptive filtering with synthetic adaptive factor for GPS positioning in road information updating in GIS

The project for the road information updating in geographical information system by using kinematic GPS method has been finished for county and country roads in China. The kinematic positioning results of commercial GPS navigation software, differential GPS positioning and an adaptively robust filtering are compared and analyzed. A synthetic adaptive factor by combining two kinds of adaptive factors is proposed for adjusting the contributions of kinematic model information and measurements on the state estimates, one is constructed with the statistics of discrepancy of kinematic model predicted state and estimated state from the measurements, and the other is set up by using the statistic of predicted residual vector. It is shown by experiments that the adaptively robust filtering with synthetic adaptive factor is valid in the cases with or without adequate GPS measurements. The calculation procedure is similar to the standard Kalman filter and navigation results are robust in controlling the influences of the outliers of the GPS measurements and kinematic state disturbing of the vehicle. The accuracy of adaptively robust filtering with only the GPS pseudo-ranges can meet the requirements of the road information updating for 1:250000 digital maps. © 2008 SPIE.