Comparación de técnicas de detección para sistemas MIMO masivo

Abstract

Massive MIMO systems are a technology that helps 5G to provide greater speed and efficiency of data transfer in cellular networks due to the large number of antennas that use in base stations. This increase brings a disadvantage when it comes to detecting signals at the base station because it increases noise and interference. Research has evaluated many detectors looking for those with the best performance and low computational complexity, one of the detectors is the successive noise cancellation detectors, the strongest signal is detected, and then its interference is subtracted from the received signal to later detect the following signs. This investigation compares the efficiency and computational complexity of the SIC (successive interference cancellation) and OSIC (ordered successive interference cancellation) detectors, based on the MMSE (minimum mean square error) and ZF (zero forced) detectors to detect the signal, and then cancel its interference from the received signal. The OSIC detector orders the signal based on SNR (signal-to-noise ratio) or SINR (signal-to-interference ratio plus noise), detecting the signal with the highest SNR or SINR and then proceeding to successive interference cancellation. So, some scenarios of a massive MIMO system are simulated in Matlab, with 16 transmitting antennas (mobile device) and 64, 100 and 200 receiving antennas (base station) using 16, 64 and 128 QAM modulation to evaluate efficiency and computational complexity of each detector proposed. Finally, the results are presented using graphs of SNR versus BER (bit error rate) to evaluate the efficiency of the ZF-SIC, MMSE-SIC, OSIC-SINR and OSIC-SNR detectors in each QAM modulation for each massive MIMO system dimension simulated, and a graph evaluating the convergence time of each detector to evaluate its computational complexity.