Networked sensing technique for 6G integrated sensing and communication

Abstract

Recently, integrated sensing and communication (ISAC) has been listed as one of the six key usage scenarios in the sixth-generation (6G) network by International Telecommunication Union (ITU). Specifically, base stations (BSs) in 6G network will emit orthogonal frequency division multiplexing (OFDM) signals not only to convey information to communication users, but also to sense the environment with ultra-high range/angle resolutions. The key challenge of 6G ISAC lies in how to achieve high-resolution sensing in a communication network. In this thesis, we aim to leverage the technique of networked sensing to tackle the above challenge. Specifically, in practice, the BSs can fuse their observations to reap the joint estimation gain. We will propose various signal processing techniques to enable network sensing for 6G ISAC. The first work considers an OFDM-based multi-cell localization system, where multiple BSs emit downlink OFDM signals simultaneously to localize the targets. A novel two-phase sensing framework is proposed, under which we design a model-free range estimation approach by leveraging the OFDM channel estimation technique for determining the delay values of all the two-way BS-target-BS paths in Phase I and localize each target based on its distances to various BSs in Phase II. Especially, we build new theory about data association in Phase II, which is a long-standing issue for multi-anchor multi-target localization. Our second work extends the networked sensing results from a pure line-of-sight (LOS) environment to a complicated multipath environment. We propose efficient methods that can mitigate the effect arising from non-line-of-sight (NLOS) paths on 6G multi-cell localization. In our third work, we extend our study from localization to target three-dimension (3D) reconstruction. We consider a multi-view approach, under which a mobile transceiver equipped with multiple transmit and receive antennas moves to different known sites to image an extended target from different angles, and fuse all the single-view images into a multi-view image at last. It is shown that the multi-view technique can significantly improve the reconstruction quality. Last, we build a millimeter wave (mmWave) ISAC platform that works at 27-29 GHz. We implement our proposed methods on this platform and achieve very high localization accuracy. In summary, in this thesis, we propose advanced signal processing technique to enable point target localization and extended target reconstruction via the networked sensing technique in 6G ISAC systems. Platform is also built to verify our results. These works can provide new insight into investigation of the 6G ISAC technique.