Energy-Efficient and Low-Complexity Spatial Modulation System with Deep Neural Network-Based Capacity-Optimized Antenna Selection

In this paper, a new telecommunication system with high data rate, energy efficiency, and low complexity is proposed by combining COAS and spatial modulation (SM) using a DNN. The proposed technique is the integration of DNN, COAS and SM and is referred to as DNN-COAS-SM (DCS). Spatial modulation improves data transmission efficiency by encoding additional information using the indices of active antennas. The proposed DCS system improves system performance and reduces complexity by using a DNN-based COAS technique to optimize antenna selections. The integration of DNN into the antenna selection process improves the speed of decision-making, ensuring high performance even in complex communication environments. Evaluated in Rayleigh fading channels using M-QAM, this system shows improvement in bit error rate performance compared to conventional SM systems. Moreover, since DCS is DNN-based, the proposed system requires less computational overhead than conventional COAS systems