A multi-rate multiple model track-before-detect particle filter

Current particle filter track-before-detect (PF-TBD) algorithms assume constant velocity (CV) motion and the filter updates at a full-rate (i.e. at every measurement scan). Previous work in multi-rate processing, via a discrete wavelet transform (DWT), has shown that multi-rate tracking algorithms can provide comparable performance at a lower computational cost. To date, these multi-rate approaches have not yet been applied to low signal-to-noise ratio (SNR) targets. This paper presents a full-rate multiple model particle filter for track-before-detect (MMPF-TBD) and a multi-rate multiple model track-before-detect particle filter (MRMMPF-TBD) that extends the areas mentioned above and tracks low SNR targets which perform small maneuvers. The MRMMPF-TBD and MMPF-TBD both use a combined probabilistic data association (PDA) and maximum likelihood (ML) approach. The MRMMPF-TBD provides equivalent root-mean-square error (RMSE) performance at substantially lower particle counts than a full-rate MMPF-TBD. A performance analysis for various SNR and particle count scenarios is presented.