TY - GEN
T1 - Joint Vehicular Social Semantic Extraction, Transmission and Cache for High QoE Digital Twin
AU - Ren, Xintian
AU - Wu, Jun
AU - Pan, Qianqian
AU - Mumtaz, Shahid
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - Digital twins have been extensively explored in vehicular social networks, while wireless communication quality is limited to support digital twins and their applications due to the high mobility of the vehicular environment. To address this issue, we propose a semantic communication empowered two-level digital twin architecture, called Semantic Twin, which supports reliable communication of efficient cloud-edge collaborative vehicular digital twin, specifically comprising low-level semantic twin (L-SemTwin) and high-level semantic twin (H-SemTwin). First, we design a social behavior semantic extraction scheme based on semantic encoder on the vehicle side to capture essential content features. Second, a semantic transmission scheme in vehicle-to-everything communication is performed to reduce the overall transmission burden and error rate, and build L-SemTwin. Moreover, we propose a deep reinforcement learning-based semantic caching strategy with the assistance of city-wise semantic information in cloud-side H-SemTwin. The experiment results demonstrate the promotion under the proposed architecture compared to conventional methods in terms of the quality of communication and user experience in vehicular social networks.
AB - Digital twins have been extensively explored in vehicular social networks, while wireless communication quality is limited to support digital twins and their applications due to the high mobility of the vehicular environment. To address this issue, we propose a semantic communication empowered two-level digital twin architecture, called Semantic Twin, which supports reliable communication of efficient cloud-edge collaborative vehicular digital twin, specifically comprising low-level semantic twin (L-SemTwin) and high-level semantic twin (H-SemTwin). First, we design a social behavior semantic extraction scheme based on semantic encoder on the vehicle side to capture essential content features. Second, a semantic transmission scheme in vehicle-to-everything communication is performed to reduce the overall transmission burden and error rate, and build L-SemTwin. Moreover, we propose a deep reinforcement learning-based semantic caching strategy with the assistance of city-wise semantic information in cloud-side H-SemTwin. The experiment results demonstrate the promotion under the proposed architecture compared to conventional methods in terms of the quality of communication and user experience in vehicular social networks.
KW - Deep Reinforcement Learning
KW - Digital Twin
KW - Semantic Communication
KW - Vehicular Social Network
UR - http://www.scopus.com/inward/record.url?scp=85187336173&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85187336173&partnerID=8YFLogxK
U2 - 10.1109/GLOBECOM54140.2023.10437533
DO - 10.1109/GLOBECOM54140.2023.10437533
M3 - Conference contribution
AN - SCOPUS:85187336173
T3 - Proceedings - IEEE Global Communications Conference, GLOBECOM
SP - 2275
EP - 2280
BT - GLOBECOM 2023 - 2023 IEEE Global Communications Conference
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2023 IEEE Global Communications Conference, GLOBECOM 2023
Y2 - 4 December 2023 through 8 December 2023
ER -