The sixth generation (6G) systems will likely employ orthogonal frequency division multiplexing (OFDM) waveform for performing the joint task of sensing and communication. In this paper, we design an OFDM system for integrated sensing and communication (ISAC) and propose a novel approach for passive target detection in an indoor deployment using a data driven AI approach. The delay-Doppler profile (DDP) and power delay profile (PDP) is used to train the proposed AI-based detector. We analyze the detection performance of the proposed methods under line of sight (LOS) and non-line of sight (NLOS) conditions for various training strategies. We show that the proposed method provides 10 dB performance improvement over the baseline for 80% target detection under LOS conditions and the performance drops by 10-20 dB for NLOS depending on the usecase scenarios.

翻译：第六代（6G）系统可能采用正交频分复用（OFDM）波形来执行感知与通信的联合任务。本文设计了一种用于集成感知与通信（ISAC）的OFDM系统，并提出了一种基于数据驱动人工智能方法的室内部署被动目标检测新方案。利用延迟-多普勒分布（DDP）和功率延迟分布（PDP）来训练所提出的基于AI的检测器。我们分析了所提方法在视距（LOS）和非视距（NLOS）条件下针对不同训练策略的检测性能。结果表明，在LOS条件下，所提方法在80%目标检测率下相较于基线方法性能提升10dB；而在NLOS条件下，根据具体应用场景，性能下降10-20dB。