Answers are grounded in this paper and cite their sources.
In this letter, a residual attention convolution (RAC) based signal detection algorithm is presented for orthogonal frequency division multiplexing with index modulation (OFDM-IM) to enhance bit-error-rate (BER) performance with low computation time. In the RAC-IM, the received signal is preprocessed into polar coordinates to ensure the network detect the index information first. Then, the processed data is input to the feature extraction network. A residual attention module is then introduced, utilizing a residual structure with skip connections to accelerate training and enhance detection accuracy. Additionally, an effective channel attention (ECA) mechanism is integrated with the residual framework to evaluate the channel importance and assigns the vital channels with larger weights to extract the important index and carrier features. Finally, the output module outputs probability and a hard decision is utilized to recover the data bits. Simulation results demonstrate the proposed algorithm can achieve better BER performance with low computation time.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: