A particle filter to mitigate jamming for GPS navigation

This paper studies a deterministic particle filter for detecting and estimating jamming for GPS navigation. Jamming results in an sudden increase of the GPS measurement noise variance, hence a degradation of the positioning solution. The proposed methodology extends the state vector by including both the unknown noise variances and a latent process indicating variance jumps. A fixed lag smoothing algorithm makes the detection more robust to outliers. The sparseness of variance changes is advantageously taken into account in the proposed strategy