Little Pilot is Needed for Channel Estimation with Integrated Super-Resolution Sensing and Communication

• Published in: 2024 IEEE Wireless Communications and Networking Conference (WCNC) pp. 1 - 6
• Main Authors: Xu, Jingran, Wang, Huizhi, Zeng, Yong, Xu, Xiaoli
• Format: Conference Proceeding
• Language: English
• Published: IEEE 21.04.2024
• Subjects: Accuracy; Array signal processing; Channel estimation; Estimation; Simulation; Superresolution; Training
• ISSN: 1558-2612
• DOI: 10.1109/WCNC57260.2024.10570698

Integrated super-resolution sensing and communication (ISSAC) is a promising technology to achieve extremely high sensing performance for critical parameters, such as the angles of the wireless channels. In this paper, we propose an ISSAC-based channel estimation method, which requires little or even no pilot, yet still achieves accurate channel state information (CSI) estimation. The key idea is to exploit the fact that subspace-based super-resolution algorithms such as multiple signal classification (MUSIC) do not require a priori known pilots for accurate parameter estimation. Therefore, in the proposed method, the angles of the multi-path channel components are first estimated in a pilot-free manner while communication data symbols are sent. After that, the multi-path channel coefficients are estimated, where very little pilots are needed. The reasons are two folds. First, compared to the conventional channel estimation methods purely relying on channel training, much fewer parameters need to be estimated once the multi-path angles are accurately estimated. Besides, with angles obtained, the beamforming gain is also enjoyed when pilots are sent to estimate the channel path gains. To rigorously study the performance of the proposed method, we first consider the basic line-of-sight (LoS) channel. By analyzing the minimum mean square error (MMSE) of channel estimation and the resulting beamforming gains, we show that our proposed method significantly outperforms the conventional methods purely based on channel training. We then extend the study to the more general multipath channels. Simulation results are provided to demonstrate our theoretical results.