Experimental Evaluation of Machine learning based MIMO-OFDM System for Optimal PAPR and BER

Abstract

The hypothetically convenient structure is the Multiple-Input Multiple-Output Orthogonal Frequency Division Multiplexing (MIMO-OFDM) technique that is employed for upcoming generations in wireless communication systems. Some of the benefits offered by MIMO-OFDM are enhanced spatial multiplexing, reliability and network throughput, and so on. Due to the integration of spatial antenna that is based on multi-stream, the problems which are related to significantly high power takes place in the system of OFDM and provides complex processing strategies. Some of the popularly known systems that are used for standardizing the Peak to average power ratio (PAPR) are partial transmit sequences (PTS), adoptive tone reservation (ATR), probabilistic mapping, and clipping which are required to be truncated and aims for minimizing the operational cost. The framework of hybrid Selective Mapping (SLM)-PTS proposed in this paper minimizes the operational cost by integrating strategies of PTS and SLM. A reduction approach that is suitable for PAPR and BER are chosen for optimization purposes depending on the statistical threshold constraint of PAPR and Bit Error Rate (BER). Thus, the system preferred with the help of the machine learning technique demonstrates the efficiency in implementing a generalized strategy to evaluate a low complexity MIMO-OFDM model. Ultimately, with the help of the PAPR and BER techniques-driven from value bound the performance of the error rate is evaluated in this framework that interactively changes from one technique.