Fine-Grained Location Privacy Protection in 6G ISAC Systems Using Deep Learning

Integrated sensing and communication (ISAC) is a pivot of 6th generation wireless networks. ISAC addresses spectrum congestion and reduces hardware costs with applications in localization, tracking, drone management, and human activity recognition [1]. However, the broadcast nature of wireless transmission introduces privacy and security challenges. To address these, we propose a novel scheme using channel obfuscation to create dummy locations, thus controlling location detection resolution. By leveraging a deep neural network (DNN) integrated with a customized genetic algorithm (GA), our approach efficiently generates subcarrier obfuscating vectors, solving the high-dimensional constrained optimization problem. Experimental results demonstrate a 97% success rate in location privacy protection, marking a significant advancement in ISAC systems fine-grained location privacy.