Number Detection for Backscatter Communication Systems: Identification of the Number of Active and Inactive Tags

Multiple tag signal detection is of great importance to the backscatter communication. However, existing works implicitly assume that all tags are in their active modes, which is not always true in practice particularly when there are some tags that cannot work or are in the sleep mode. To fill in this gap, in this paper, we consider the detection of active and inactive tags for backscatter communications, and propose a novel deep learning model named number detection network (NdNet) by exploiting an advanced spatial gating multilayer perceptron (SGMLP) architecture. Through comprehensive simulations, we demonstrate that our proposed SGMLP block markedly outperforms traditional MLP architectures. Additionally, the NdNet shows a superior performance in detecting the number of active and inactive tags compared to conventional deep learning models.