Answers are grounded in this paper and cite their sources.
This paper proposes a method for classifying combinations of instructions EMI of an embedded system by using a convolutional neural network. First, measuring the EMI emitted by 16 different combinations of instructions from an embedded system, then transform the EMI data to spectrogram and build a massive volume dataset to train the convolutional neural network. By analyzing the confusion matrix, we classify those similar combinations of instructions as the same groups. After several interactions, the 16 combinations of instructions can be classified into 5 groups. In the final result, the convolutional neural network model has 98% training accuracy and 83% prediction accuracy.
No takes yet. Share an insight, caveat, or question.
Yuan et al. (2019) studied this question.
Synapse has enriched one closely related paper. Consider it for comparative context: