Artificial Neural Network (ANN) based Classification of High and Low PAPR OFDM signals

In the upcoming era, 5G is expected to enhance communications in terms of high-speed data rate, low latency, and efficient power utilization. Orthogonal frequency division multiplexing (OFDM) is the modulation technique of 5G. However, OFDM suffers from a large peak-to-average power ratio (PAPR). This degrades the performance of the power amplifiers (PA). As a result, the PAPR deteriorates the overall power efficiency of the 5G system. So, it is often highly essential to find which type of OFDM signals results in high PAPR. This classification will enable PAPR reduction techniques such as Clipped-Control Tone Reservation (CC- TR) to focus effectively on the classes of OFDM signals that have high PAPRproblems. So, we have proposed a model of Artificial Neural Network (ANN) for classifying OFDM signals having high and low PAPR. Here we have compared our ANN's performance, by using different types of activation functions, optimizers, signal chunk samples, and input data types such as combinations of bits and complex symbols to demonstrate the varying estimation ability and accuracy of our ANN model.