In this work, the performance of logt-, GMOS(Geometric Mean Order Statistic), TMOS(Trimmed MOS) and IE-CFAR (Inclusion/Exclusion) detectors are investigated in presence of log-normal and K distributed clutter. First, for a finite number of clutter samples, dependence of the false alarm probability P FA upon clutter parameters is examined. The CFAR property for the case of log-normal clutter is maintained while the P FA depends somewhat on the shape parameter of the K distribution. Then, by carrying out Monte-Carlo simulations, we show that for the case of log-normal clutter a small detection difference exists between the underlying CFAR detectors. In the case of K-distributed clutter, there is a significant detection difference for small values of the shape parameter (spiky clutter case).
This paper concerns the problem of estimating the parameters of the K plus noise distribution. In a previous work, it has been shown that, in the multilook scenario, the modified fractional order moment estimator (MFOME) has about the same estimation accuracy as the [zlog(z)] method, but lower computational complexity. However, significant estimation errors have been observed in the single look scenario, low sample size, and large values of the K-distribution shape parameter. Moreover, the computational complexity of the [zlog(z)] estimator discourages its implementation in practical applications. The aim of this work is to estimate the shape parameter of the K-distribution with reduced computational complexity. The problem can be formulated as a supervised many-to-one sequence prediction. We propose here a hybrid model including convolutional and long-short-term-memory (LSTM) neural networks (NN). Estimation performance is investigated by processing both simulated and real clutter data.
Parameter estimation of radar clutter is considered as a critical task for the development of target detectors. This work covers the shape parameter estimation of K-clutter plus noise using a modified fractional order moments based approach (MFOME). Closed form of the FOME with fixed fractional order moment is derived in a previous work [11] where undesirable estimation errors are produced in some cases with single look data and low sample sizes. In order to achieve better estimation performance, the fractional order moment and the shape parameter should be optimized together. To this effect, a numerical formula of the corresponding fitness function is given and unconstrained nonlinear optimization method based on the Nelder-Mead simplex algorithm is used to compute the unknown parameters. Via simulated K-clutter plus noise data, the effect of the fractional order on the estimation accuracy is studied firstly. Then, comparisons with existing HOME, FOME and [zlog(z)] methods are conducted to illustrate the efficiency of the proposed estimator.
30 June 2020 in Ingénierie des Systèmes d'information