Optimized transmit antenna selection and self-attention based convolutional resource allocation model for massive MIMO technology

Abstract

Massive multiple input and multiple output (MIMO) plays an important role in enhancing the transmission reliability and capacity of the transmitting channel. However, the resource allocation (RA) and transmit antenna selection (TAS) scheme are essential for massive MIMO systems to reduce implementation costs and complex operations. Because the presence of a large antenna consumes enormous resources that must be reduced in order to develop a user-friendly model. Hence, this article introduces a novel TAS and RA scheme in a massive MIMO system that can enhance communication performance efficiently. In this study, an improved sheep flock optimization algorithm (ISFOA) is first emphasized to select the efficient antenna, thus effectively minimizing the cost and the design complexity. Then, a novel deep learning (DL) based self-attention aided deep convolutional neural network (SA-DCNN) model for the stable allocation of resources to all available users is proposed. Thus, the user equipment (UE) can develop a better quality path and reduce the power consumption of the entire system. In the experimental scenario, the performance of the proposed model is compared with an existing technique based on sum rate, spectral efficiency (SE) and energy efficiency (EE) by varying base station (BS) antennas, varying user, signal to noise ratio (SNR), and sub-carriers along with the computational performance. Particularly, when the SNR=10 dB, the proposed model obtains the SE of about 50 bits/s Hz compared to the existing techniques.