Multiple target track-before-detect in compound Gaussian clutter

In this paper, we extend the multiple transition mode track- before-detect (TBD) algorithm to track multiple low observable targets in compound Gaussian sea clutter. The proposed TBD framework uses the un-thresholded fast time radar measurements to track multiple targets in low signal-to-clutter ratios (SCRs). The TBD is implemented using particle filtering (PF), and we derive the generalized likelihood ratio needed to update the particle weights. The maximum likelihood estimate of the texture and the covariance matrix of the speckle are also derived and implemented using a fixed point algorithm. The tracking performance of the proposed algorithm is investigated using three low observable targets that enter and leave the field of view (FOV) at different time steps and under varying environmental conditions.