Machine Learning-Aided ISAC with OTFS for 6G

Integrated sensing and communications (ISAC) is set to enable a range of new services in future Sixth Generation (6G) networks. To support simultaneous sensing and broadband communications, the physical layer (PHY) must be designed to operate under challenging scenarios. Orthogonal time-frequency space (OTFS) is a strong 6G waveform candidate due to its robustness to high mobility and seamless sensing integration. However, its high overhead remains a challenge. To address this, we are investigating a novel low-complexity, zero-overhead channel estimation method for OTFS, leveraging machine learning (ML) to enhance performance and advance towards an intelligent PHY.