A channel state prediction for multi-secondary users in a cognitive radio based on neural network
Nakisa Shamsi, Amir Mousavinia, Hadi Amirpour · 2013
Sensing the spectrum and accessing it are two important challenges for any secondary user who wants to use the available communication channel in a cognitive radio system. This spectrum utilization can be improved by secondary users through using free licensed channels in the absence of the primary user. In this work, we seek two objectives, channel estimation in predictive modeling scenario and multi-secondary user scenario using Artificial Neural Networks. Time Delay Neural Network (TDNN) and Recurrent Neural Network (RNN) have been selected to design the predictor. The accuracy of this forecasting can easily improve the spectrum utilization. In the second scenario, a channel status predictor is configured for each secondary user enabling them to identify the best available channel. Simulation results show that the prediction error has been reduced to less than 14% in average. However, in some cases it can predict the next channel status correctly with zero error prediction.