Research on OFDM-based Integrated Sensing and Communication Method

In this paper, we utilize OFDM signals and deep learning technology to achieve environment sensing through narrow bandwidth OFDM signal, enabling the perception of personnel status and position in enclosed spaces. Firstly, the paper provides a detailed introduction to the characteristics and application fields of OFDM signals, along with an in-depth analysis of associated issues. Second, we introduce the error vector spectrum (EVS) method for accurate channel quality evaluation through spectral analysis of error between received and real signals, thereby enhancing channel estimation accuracy. The DNN model is trained to predict user location by transmitting narrow bandwidth OFDM signals indoors and collecting data at the receiving end. Experimental results demonstrate significant accurate in indoor positioning.