Adaptive Channel Estimation based on Deep Learning

Abdul Karim Gizzini (1) , Marwa Chafii (1) , Ahmad Nimr (2) , Gerhard Fettweis (2)

1 ETIS - UMR 8051 - Equipes Traitement de l'Information et Systèmes
2 TU Dresden - Technische Universität Dresden = Dresden University of Technology

Channel state information is very critical in various applications such as physical layer security, indoor localization, and channel equalization. In this paper, we propose an adaptive channel estimation based on deep learning that assumes the signal-to-noise power ratio (SNR) knowledge at the receiver, and we show that the proposed scheme highly outperforms linear minimum mean square error based channel estimation in terms of normalized minimum square error, with similar order of online computational complexity. The proposed channel estimation scheme is also evaluated for an imperfect estimation of the SNR and showed to be robust for a high SNR estimation error.

Mots clés: Adaptive channel estimation; Deep learning; Machine learning; LMMSE

Domaines: Traitement du signal et de l'image [eess.SP]; Intelligence artificielle [cs.AI]; Apprentissage [cs.LG]; Performance et fiabilité [cs.PF]; Complexité [cs.CC]

Abdul Karim Gizzini, Marwa Chafii, Ahmad Nimr, Gerhard Fettweis. Adaptive Channel Estimation based on Deep Learning. The 2020 IEEE 92nd Vehicular Technology Conference: VTC2020-Fall, Oct 2020, Victoria, Canada. ⟨hal-02895538⟩