Sinir ağına dayalı görünür ışık iletişim kanalı tahmini

Visible light communication techniques are increasingly emerging to address these needs with growing criteria for easy and secure communication between devices. One of the main strategies using many links for high-performance, stable communication is multiple input and output (MIMO) systems. But the increased number of communication links adds to the difficulty of the channel calculation, which is necessary if transmitted data is to be correctly decoded. Therefore, it is important to improve detailed and effective methods of channel estimation. We disclose the performance of neural network channel estimation approaches to boost the performance of MIMO visible light channel estimation. The proposed estimation method compared with linear regression to check the accuracy of the proposed method. Simulation results of both NN and linear regression method compared and it's clear that the BER was in range of 10-5 for the NN which is better than the BER for linear regression method which was in range of 10-2. The main advantage of using NN for VLC channel estimation is the reduction of SNR while keeping low BER. The estimation results show that the presented algorithm is comparable to the MATLAB standard channel model. The results confirm that this procedure can be used efficiently in channel estimation. Keywords: Channel estimation, Neural networks, MIMO, visible light communications, SNR, BER