Analysis on the Channel Prediction Accuracy of Deep Learning-based Approach

In recent days, the vehicular communication system (VCS) plays an important role in driving safety and traffic information. In VCS, one of the most important factors that affects the system performance is the channel prediction. The accurate channel prediction is a necessary part for secure vehicle-to-vehicle communication. The channel prediction in VCS has many challenges and these challenges reduce VCS performance. In this paper, we analyze the impact of the deep learning-based channel prediction algorithm for vehicle-to-vehicle communication to improve the channel prediction accuracy of VCS. We consider the algorithm called channel adaptive transmission (CAT) which uses the long short-term memory (LSTM) networks for channel prediction. The proposed approach achieves 2.6 dBm of root mean square error and over 97% of prediction accuracy. The result shows that this algorithm can be utilized efficiently in channel prediction.