An optimal filter to reduce BER utilizing RLS and firefly

Abstract

Data communication network suffers due to symbol interference and un-optimized channel response. In recent years faster communication architectures like OFDM were designed to transmit the data as fast as and compatible with modern day communication devices. In order to utilize the channel efficiently, data should be filtered and precise. Swarm Intelligence based recursive least square algorithm has been developed utilizing Extended Firefly Algorithm for the geometric transformation of the data. The optimized filtered data has been cross-verified using Support Vector Machine approach. The permutation matrix of proposed work has been compared with results obtained using Kalman filtering. Results demonstrate that if a filter is designed significantly relative to the single objective function of the optimization algorithm, it generates quite good estimates. The performance of the proposed structure is evaluated using Bit Error Rate and Average Logarithmic Error measures.