HYBRID GA-PSO TRAINED FUNCTIONAL LINK ARTIFICIAL NEURAL NETWORK BASED CHANNEL EQUALIZER

Channel equalization is an important field of adaptive signal processing. When significant noise is added to the transmitted signal in the channel, the received signal at each instant can be considered as a nonlinear function of the past values of transmitted signal. The overall channel response becomes a non-linear dynamic mapping problem. Hence, the channel needs to be equalized using best of the non-linear approximators. In this paper, Functional Link Artificial Neural Network is used as equalizer by training with Hybrid GA-PSO Algorithm as the Least Mean Square (LMS) methodology is not being able to meet the requirements under noisy conditions. From the simulations and results it can be seen that proposed Hybrid GA-PSO training methodology can be considered as a better training algorithm compared to previously proposed GA and PSO trained FLANN.