TY - GEN
T1 - AirJSCC
T2 - 2023 IEEE Global Communications Conference, GLOBECOM 2023
AU - Hui, Yingzhe
AU - Chen, Shuyi
AU - Meng, Weixiao
AU - Chen, Hsiao Hwa
AU - Sun, Wen Bin
AU - Ma, Lin
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - Reconfigurable intelligent surface (RIS) is a programmable metasurface composed of sub-wavelength meta-atoms and a controller. Recent research works have shown that multi-layer RISs can form over-the-air neural networks. This enables support for a variety of tasks, including image recognition, mobile communication coding-decoding, and real-time multi-beam focusing. In this work, we present a RISs-based over-the-air joint source channel coding (JSCC) scheme, named as AirJSCC, where multi-layer RISs are used to realize wave-based complex-valued neural networks (CVNNs). AirJSCC transforms the computation process inherited in JSCC into the transmission of wireless signals through RISs, and thus many benefits, such as light-speed computation and high parallel processing capability, can be achieved. To the best of our knowledge, this is the first deep JSCC scheme that incorporates RISs and CVNNs. Simulation results demonstrate that AirJSCC achieves better image reconstruction performance compared with the baseline scheme, even under low signal-to-noise ratio (SNR) and limited bandwidth, and exhibits robustness against varying channel conditions.
AB - Reconfigurable intelligent surface (RIS) is a programmable metasurface composed of sub-wavelength meta-atoms and a controller. Recent research works have shown that multi-layer RISs can form over-the-air neural networks. This enables support for a variety of tasks, including image recognition, mobile communication coding-decoding, and real-time multi-beam focusing. In this work, we present a RISs-based over-the-air joint source channel coding (JSCC) scheme, named as AirJSCC, where multi-layer RISs are used to realize wave-based complex-valued neural networks (CVNNs). AirJSCC transforms the computation process inherited in JSCC into the transmission of wireless signals through RISs, and thus many benefits, such as light-speed computation and high parallel processing capability, can be achieved. To the best of our knowledge, this is the first deep JSCC scheme that incorporates RISs and CVNNs. Simulation results demonstrate that AirJSCC achieves better image reconstruction performance compared with the baseline scheme, even under low signal-to-noise ratio (SNR) and limited bandwidth, and exhibits robustness against varying channel conditions.
UR - http://www.scopus.com/inward/record.url?scp=85187375425&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85187375425&partnerID=8YFLogxK
U2 - 10.1109/GLOBECOM54140.2023.10437941
DO - 10.1109/GLOBECOM54140.2023.10437941
M3 - Conference contribution
AN - SCOPUS:85187375425
T3 - Proceedings - IEEE Global Communications Conference, GLOBECOM
SP - 4092
EP - 4097
BT - GLOBECOM 2023 - 2023 IEEE Global Communications Conference
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 4 December 2023 through 8 December 2023
ER -