DECODING OF TURBO CODE AND POLAR CODE USING DEEP LEARNING FOR VISIBLE LIGHT COMMUNICATION

The success of neural network-based systems in areas such as image classification, computer vision, pattern recognition, Natural language Processing etc. has made its usage also expand in the area of Wireless communication. This paper investigates the decoding of two codes widely used in modern communication viz, Turbo Codes and Polar Codes using Deep Learning (DL) methods. The aim of this study is to explore the feasibility of using DL architectures based on Deep Neural Networks (DNN) and Recurrent Neural Networks (RNN) for decoding of Polar Codes and Turbo Codes, respectively. The decoding performance of DNN based Polar Codes is also investigated based on number of neurons in each layer, activation functions and number of layers. Simulation is carried out in MATLAB for the communication system independently for the above codes over an Additive White Gaussian Noise-Visible Light Communication Channel employing Colour Intensity Modulation. The results compare the performance of the traditional decoding algorithm with the proposed DL approach over different Signal to Noise Ratio (SNR) regimes. The RNN based Turbo decoder outperforms the conventional Turbo decoder in terms of Bit Error Rate performance at lower SNRs and the DNN based Polar decoder has a similar performance as of the conventional Polar decoder.