A BAYESIAN APPROACH FOR LOW PAPR OFDM MODULATION IN MASSIVE MIMO DOWNLINK SYSTEMS WITH MULTIUSER INTERFERENCE CANCELLATION

The impact of communication has greatly risen with contemporary technology. The communication parameters have improved as a result of recent advancements in the communication system. By using multipath propagation and numerous transmit and receive antennas, multiple-input multiple-output, or MIMO, is a methodology used in radio to increase the capacity of a radio link. The problem of declining peakto-average power ratios (PAPR) in massive multiple-input multiple-output (MIMO) downlink systems relied upon orthogonal frequency-division multiplexing (OFDM) is taken into consideration. To be more precise, the challenge is to identify an OFDMmodulated signal with a low PAPR and the ability to simultaneously enable multiuser interference (MUI) cancellation provided a set of symbol vectors to be sent to K users. Contrary to earlier research that used convex optimisation to address the issue, we use a Bayesian strategy and create an effective PAPR reduction approach by taking advantage of the transmit array's redundant degrees of freedom. A low PAPR signal with the majority of its samples focused on the borders may result from treating the desired signal as a random vector with a hierarchical truncated Gaussian mixture prior. The signal and estimates of the prior model's hyper-parameters are obtained using a variational expectation-maximization (EM) technique. The variational EM framework also incorporates the generalised approximation message passing (GAMP), which significantly reduces the computing cost of the suggested solution. According to simulation findings, our suggested technique outperforms existing methods significantly with regards to both PAPR reduction and computing complexity.