Mutual Localization and Positioning of Vehicles Sharing GNSS Pseudoranges: Sequential Bayesian Approach and Experiments

In many cooperative Intelligent Transportation Systems (ITS) applications, absolute positioning and relative localization are key issues. When vehicles share GNSS positions, there are often non negligible common-mode errors due mainly to GNSS signal-in-space. Cooperative observation techniques allow estimating common biases on the measured pseudodistances to correct these errors and to increase absolute positioning and relative localization accuracy. After having studied some structural properties of the problem in its general form, a low computational cooperative tightly-coupled approach is proposed using sequential Kalman filtering and convex data fusion. As a case study, we consider two vehicles which cooperate and exchange information in such a way that each vehicle can track the partner's position and improves its absolute position by merging common biases estimates. Experimental results are presented to illustrate the performance of the proposed approach in comparison with a classic standalone method.