Artificial neural network based channel equalization using battle royale optimization algorithm with different initialization strategies

In digital communication, the transmitted data is affected due to channel distortion. In terms of Inter-Symbol Interference (ISI), the distortion is occurs due to the dispersive nature of channel. The channel equalization technique is used at the receiver end for their reliability and high-speed communication by reducing the effects of ISI. In this paper an effective equalizer based on Artificial Neural Network (ANN) is proposed. The weights of ANN are trained by proposed Battle Royale Optimization (BRO). The objective function of ANN-BRO based equalizer is to minimize Mean Square Error (MSE) values where the error value is estimated based on the transmitted signal and the equalizer output. The BRO is modified by different initialization methods like Random Number Generation (RNG), Opposition based learning (OBL), Quasi-Opposition based learning (QBL), Tent Map and chaotic methods. The experimental results of the proposed equalizer is evaluated and compared with various initialization and optimization approaches. The performance measures such as MSE, Mean Square of the Residual Error (MSRE), and Bit Error Rate (BER) are evaluated to show the efficiency of the proposed method.