Providing URLLC Service in Multi-STAR-RIS Assisted and Full-Duplex Cellular Wireless Systems: A Meta-Learning Approach
Eghbali, Yasoub; Taskou, Shiva Kazemi; Mili, Mohammad Robat; Rasti, Mehdi; Hossain, Ekram (2024-01-03)
IEEE
03.01.2024
Y. Eghbali, S. K. Taskou, M. R. Mili, M. Rasti and E. Hossain, "Providing URLLC Service in Multi-STAR-RIS Assisted and Full-Duplex Cellular Wireless Systems: A Meta-Learning Approach," in IEEE Communications Letters, vol. 28, no. 3, pp. 513-517, March 2024, doi: 10.1109/LCOMM.2023.3349377.
https://creativecommons.org/licenses/by/4.0/
© 2024 The Authors. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
https://creativecommons.org/licenses/by/4.0/
© 2024 The Authors. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
https://creativecommons.org/licenses/by/4.0/
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:oulu-202402211913
https://urn.fi/URN:NBN:fi:oulu-202402211913
Tiivistelmä
Abstract
The Simultaneously Transmitting and Reflecting Reconfigurable Intelligent Surface (STAR-RIS) technology is an innovative approach that aims to enhance the performance of sixth-generation (6G) wireless networks. This study focuses on a multi-STAR-RIS and full-duplex (FD) communication system aimed at providing ultra-reliable low-latency communication (URLLC) services. To maximize the total uplink (UL) and downlink (DL) rates, beamforming and combining vectors at the base station (BS), the transmit power of UL users, the amplitude attenuations, and phase shifts of the STAR-RISs are jointly optimized. These optimizations take into account the maximum transmit power constraints of the BS and UL users, as well as the quality of service requirements of UL and DL users. Given the non-convex nature of the optimization problem, this study proposes a novel deep reinforcement learning algorithm called Meta DDPG, which combines meta-learning and deep deterministic policy gradient. Numerical results demonstrate that a multi-STAR-RIS assisted system can obtain a higher system total rate compared to the conventional multi-RIS assisted system.
The Simultaneously Transmitting and Reflecting Reconfigurable Intelligent Surface (STAR-RIS) technology is an innovative approach that aims to enhance the performance of sixth-generation (6G) wireless networks. This study focuses on a multi-STAR-RIS and full-duplex (FD) communication system aimed at providing ultra-reliable low-latency communication (URLLC) services. To maximize the total uplink (UL) and downlink (DL) rates, beamforming and combining vectors at the base station (BS), the transmit power of UL users, the amplitude attenuations, and phase shifts of the STAR-RISs are jointly optimized. These optimizations take into account the maximum transmit power constraints of the BS and UL users, as well as the quality of service requirements of UL and DL users. Given the non-convex nature of the optimization problem, this study proposes a novel deep reinforcement learning algorithm called Meta DDPG, which combines meta-learning and deep deterministic policy gradient. Numerical results demonstrate that a multi-STAR-RIS assisted system can obtain a higher system total rate compared to the conventional multi-RIS assisted system.
Kokoelmat
- Avoin saatavuus [38840]
Ellei muuten mainita, aineiston lisenssi on https://creativecommons.org/licenses/by/4.0/