New non-linear decision feedback equalizers based on the functional-link neural network

This paper presents a new approach for the Decision Feedback Equalizer using the Functional Link Network (DFFLE) for channel equalization in digital communication systems. Results show that the new non-linear DFFLE structures offer superior performance to the Multi-Layered perceptron (MLP) based DFE and the conventional DFE-LMS which are two widely used decision feedback equalizers. The performance has been evaluated in terms of the equalizers' speed of convergence and bit error rates.<
>