A carrier-phase DGPS based V2V object sensing system using fast incremental Bayesian network

This paper describes a novel approach to moving-baseline carrier-phase differential GPS (DGPS) and a vehicle-to-vehicle object system based on the approach. In order to achieve sub-decimeter level accuracy, a Bayesian network is proposed to fuse measurements from GPS and low-cost vehicle dynamic sensors for trajectory reconstruction. A fast recursive implementation whose complexity scales linearly with the tra- jectory length is derived. Experimental results are presented to illustrate the approach's effectiveness to fuse data from GPS and vehicle dynamic sensors. To show the performance and effectiveness of the proposed vehicle-to-vehicle (V2V) object sensing system, we choose a frequency-modulated continuous wave (FMCW) radar as the benchmark for comparison.