A Whole-Net Positioning Method Based on Baseline Optimization Selection in Multi-vehicle Cooperative System

Cooperative positioning for vehicles is a typical application of Internet of Vehicle, and high-precision relative positioning in a cooperative system is a significant issue. To achieve decimeter or centimeter level accuracy, one of the main approaches is using relative positioning methods based on carrier differential global navigation satellite system (CD-GNSS). However, in urban scenarios, frequent signal outages, multipath and non-line-of-sight (NLOS) lead to lower availability and accuracy of CD-GNSS. Existing cooperative CD-GNSS schemes fuse the carrier-phase observations of different vehicles and solve all baselines in the cooperative system separately. However, it is not necessary to solve all baselines in a cooperative, and satellite observations cannot be simultaneously used when baselines are solved separately, which means that observations are not fully used. Aiming to the limitations above, we propose a whole-net cooperative positioning method based on baseline optimization selection. A baseline optimization selection scheme is introduced to search an optimal minimum baseline group to fix all vehicles. Then a whole-net positioning filter is designed to solve this baseline group using all observations in the system. An experiment involving three vehicles is conducted, and preliminary experimental results show that proposed algorithm can improve the ambiguity fixing rate and positioning accuracy in constrain environment. More significant improvement can be achieved when vehicles receive different satellites from each other.