Enhanced Truncated Statistic CFAR Based on Median Absolute Deviation and Weighted Log-likelihood Estimation

The actual working environment of radar is complex, including many interferences and non-homogeneous clutter, which seriously degrade the detection performance of radar. This paper proposes an enhanced truncated statistic CFAR (ETS-CFAR) based on median absolute deviation and weighted log-likelihood estimation. The algorithm has good detection performance in multi-target environment, and moreover, ensures good performance in clutter edge environment, which has a good ability to maintain a constant false alarm rate. The median absolute deviation is applied to improve the estimation of truncation depth, which is a robust measure and insensitive to the outlier, and the statistical interference outliers are eliminated. The weighted log-likelihood estimation is adopted to optimize the determination conditions of the weighted coefficients and improve the parameter estimation of the clutter. The proposed ETS-CFAR algorithm is evaluated in different clutter environments through simulation, and the superiority of the algorithm is explained.