Probabilistic 3D mapping based on GNSS SNR measurements

A probabilistic approach to 3-dimensional mapping is proposed that only uses data gathered by GNSS (Global Navigation Satellite System) devices. To accomplish this, the environment is gridded and a physically motivated sensor model is developed that assigns likelihoods of blockage to satellite signals based on their measured SNR (signal-to-noise ratio). It is then shown that the posterior distribution of the map represents a sparse factor graph on which a low complexity implementation of Loopy Belief Propagation can be used for efficient Bayesian estimation. Experimental results are presented which demonstrate our algorithm's ability to coarsely map in three dimensions a corner of a university campus.