WLAN Link Adaptation Based on Neural Networks

Two neural networks with two features (SNR and mutual information) are applied to link adaptation that is targeted for meeting the WLAN standard. The usefulness of neural-network-based link adaptation is validated for the WLAN and also the performance (throughput and latency) is analyzed to obtain some practical gain of link adaptation based on neural networks over the traditional approach. For the top-1 error, the accuracy is 91.2% with the multi-layer perceptron and 95% with the convolutional neural network. From the histogram distributions of mutual-information-based and SNR-based classifications, it is conjectured that the mutual information may serve more or less better as the extracted feature than the SNR for convolutional-neural-network-based link adaptation.