• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.1C
• /
• pp.33-39
• /
• 2010

Correntropy is a generalized correlation function that contains higher order moments of the probability density function (PDF) than the conventional moment expansions. The criterion maximizing cross-correntropy (MCC) of two different random variables has yielded superior performance particularly in nonlinear, non-Gaussian signal processing comparing to mean squared error criterion. In this paper we propose a new blind equalization algorithm based on cross-correntropy criterion which uses, as two variables, equalizer output PDF and Parzen PDF estimate of a set of randomly generated symbols that complies with the transmitted symbol PDF. The performance of the proposed algorithm based on MCC is compared with the Euclidian distance minimization.

• Journal of Internet Computing and Services
• /
• v.11 no.5
• /
• pp.19-26
• /
• 2010

The criterionmaximizing cross-correntropy (MCC) of two different random variables has yielded superior performance comparing to mean squared error criterion. In this paper we present a complex-valued blind equalizer algorithm for QAM and complex channel environments based on cross-correntropy criterion which uses, as two variables, equalizer output PDF and Parzen PDF estimate of a self-generated symbol set. Simulation results show significantly enhanced performance of symbol-point concentration with no phase rotation in complex-channel communication.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39A no.3
• /
• pp.149-154
• /
• 2014

The cross-correntropy concept can be expressed with inner products of two different probability density functions constructed by Gaussian-kernel density estimation methods. Blind algorithms based on the maximization of the cross-correntropy (MCC) and a symbol set of randomly generated N samples yield superior learning performance, but have a huge computational complexity in the update process at the aim of weight adjustment based on the MCC. In this paper, a method of reducing the computational complexity of the MCC algorithm that calculates recursively the gradient of the cross-correntropy is proposed. The proposed method has only O(N) operations per iteration while the conventional MCC algorithms that calculate its gradients by a block processing method has $O(N^2)$. In the simulation results, the proposed method shows the same learning performance while reducing its heavy calculation burden significantly.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.12
• /
• pp.2324-2330
• /
• 2015

In blind signal processing in impulsive noise environment the maximum cross-correntropy (MCC) algorithm shows superior performance compared to MSE-based algorithms. But optimum weight conditions of MCC algorithm and its properties related with robustness to impulsive noise have not been studied sufficiently. In this paper, through the analysis of the behavior of its optimum weight and the relationship with the MSE-based LMS algorithm, it is shown that the optimum weight of MCC and MSE-based LMS have an equal solution. Also the factor that keeps optimum weight of MCC undisturbed and stable under impulsive noise is proven to be the magnitude controlled input through simulation.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.9
• /
• pp.995-1000
• /
• 2016

The maximum cross-correntropy (MCC) algorithm with a set of random symbols keeps its optimum weights undisturbed from impulsive noise unlike MSE-based algorithms and its main factor has been known to be the input magnitude controller (IMC) that adjusts the input intensity according to error power. In this paper, a normalization of the step size of the MCC algorithm by the power of IMC output is proposed. The IMC output power is tracked recursively through a single-pole low-pass filter. In the simulation under impulsive noise with two different multipath channels, the steady state MSE and convergence speed of the proposed algorithm is found to be enhanced by about 1 dB and 500 samples, respectively, compared to the conventional MCC algorithm.