An Artificial Intelligence-Based Communication System For Processing of Modulation-Demodulation

As for the aspect of communication, in digital transmissions, the receiver side has to perform a demodulation of the modulated signals that are in use. Systems using different types of hardware are the ones that are currently in use, while communication signals come with a different mechanism of reaching their recipients. Signal classification and the subsequent demodulation are explained using a single algorithm rather than extra expenses and complicated mechanisms. This convenience is provided in this work, which tries to employ signals in modulation classification. Digital modulation signal illustrations are utilized for classification and demodulation in this study. Convolutional neural networks (CNNs), among the deep learning algorithms, have been used in recognition and classification problems. The given modular signals of four digital forms that include QPSK, QFSK, and QASK with a signal-to-noise ratio of ۰ dB, ۵ dB, ۱۰ dB, and ۱۵ dB can be utilized. Thanks to this approach, which is not dependent on the specific existing hardware, the success rate is relatively high, based on the attempts performed. It should be noted that the algorithm can be trained and tested using Python and related libraries. For demodulation procedures of these signals, the nonlinear autoregressive network with exogenous inputs (NARX) artificial neural network can be applied. The NARX method acquired the information signal with a high rate of success, nearly close to ۹۵% and due to the work done, it is possible to identify and demodulate other transmission signals without having to introduce extra hardware.

کلیدواژه ها:

Communications System ، Modulation ، Signal Classification ، Demodulation ، CNN ، Telecommunications