NONPARAMETRIC DENSITY-ESTIMATION AND DETECTION IN IMPULSIVE INTERFERENCE CHANNELS .2. DETECTORS

Nonlinear processing significantly enhances detector performance in no ngaussian noise relative to that of linear detectors. In this part of the study (Part II), several nonparametric detection schemes for impul sive noise channels are formulated using the nonparametric probability density estimators developed in Part I. The likelihood ratio test and the small-signal (locally optimum) nonlinearity provide the basis for the formulation of these nonparametric detection schemes. Several mod ifications to these basic strategies are used to compensate for inaccu racies in the density estimates. In particular, for the problem of det ecting a known signal in impulsive noise, two modifications to the sta ndard likelihood ratio test are considered: the first is adapted from robust statistics, whereas the second, the ''L1-error-based'' detector is specifically formulated for use with density estimates. Both schem es are found to perform close to the optimum likelihood ratio detector for a wide variety of impulsive noise densities (including the Class A model, the Johnson S(u) model, and the Gaussian-Laplacian mixture). From the merits of these two tests, a new detection scheme that approx imates the locally optimum nonlinearity is then developed. This detect or, which uses the nonparametric density estimators developed in Part 1, is shown to perform very well for the wide variety of impulsive and heavy-tailed densities considered in this study. This nonparametric-d ensity-estimate-based detector is also shown to outperform more conven tional nonparametric detectors in impulsive noise.