AppDAS: An Application QoS-Aware Distributed Antenna Selection for 5G Industrial Applications

• Main Authors: Onishi, Takeo, Takahashi, Eiji, Nishikawa, Yoshiaki, Maruyama, Shohei
• Format: Conference Proceeding
• Language: English
• Subjects: 5G mobile communication; Antenna selection; Deep reinforcement learning (DRL); Distributed antenna system (DAS); Industrial application; Measurement; Quality of service; Reinforcement learning; Spatial diversity; Stability analysis; Throughput
• ISSN: 2331-9860
• DOI: 10.1109/CCNC51644.2023.10059796

Next-generation wireless and mobile networks including 5G and 6G are expected to be utilized in the industrial field for applications such as the remote control of vehicles/robots thanks to their high reliability and low latency. In industrial applications, it is crucial to precisely satisfy the communication requirements of each application to ensure adequate performance in terms of safety and/or productivity. Diversity is a key feature to ensure the stability of wireless communication, and a distributed antenna system (DAS) is expected to enhance space diversity. In this paper, we propose an application-aware distributed antenna selection method for DAS to improve the performance of 5G industrial applications. To satisfy the diverse communication requirements of industrial applications, the proposed method selects the best combination of next-generation nodeB (gNB) antennas and user equipments (UEs). The enormous number of potential combinations makes it difficult to determine the optimal one by a brute-force or greedy algorithm. We therefore built a two-step selection scheme consisting of coarse and fine UE selection along with a requirement-based metric for deep reinforcement learning to solve it. End-to-end simulations to evaluate the performance of the proposed antenna selection method showed that it can satisfy 90 % of the communication requirements of applications, which is a significant improvement over the 50 % provided by the conventional approach.