Online adaptive nonlinear channel equalization using RBF neural networks

Nonlinear distortions must be compensated in many real-life systems, such as in digital satellite and microwave channels and others, so adaptive equalization is of considerable practical interest. Radial basis function neural networks have the ability to equalize nonlinear channels, and a simplified version is proposed for online implementation. We focus on the stochastic gradient technique, and get a good performance by adjusting only one coefficient and one center closest to the input vector. Simulations are included to verify the algorithm.