Abstract
Relative positioning is recognized as an important issue for vehicles in urban environ-ments.Multi-vehicle Cooperative Positioning(CP)techniques which fuse the Global Navigation Satellite System(GNSS)and inter-vehicle ranging have attracted attention in improving the perfor-mance of baseline estimation between vehicles.However,current CP methods estimate the baselines separately and ignore the interactions among the positioning information of different baselines.These interactions are called'information coupling'.In this work,we propose a new multi-vehicle precise CP framework using the coupled information in the network based on the Carrier Differential GNSS(CDGNSS)and inter-vehicle ranging.We demonstrate the benefit of the cou-pled information by deriving the Cramer-Rao Lower Bound(CRLB)of the float estimation in CP.To fully use this coupled information,we propose a Whole-Net CP(WN-CP)method which consists of the Whole-Net Extended Kalman Filter(WN-EKF)as the float estimation filter,and the Partial Baseline Fixing(PBF)as the ambiguity resolution part.The WN-EKF fuses the mea-surements of all baselines simultaneously to improve the performance of float estimation,and the PBF strategy fixes the ambiguities of the one baseline to be estimated,instead of full ambiguity resolution,to reduce the computation load of ambiguity resolution.Field tests involving four vehi-cles were conducted in urban environments.The results show that the proposed WN-CP method can achieve better performance and meanwhile maintain a low computation load compared to the existing methods.
基金项目
National Natural Science Foundation of China(61901015)
NETL
NSTL
万方数据