Our lab is dedicated to addressing various application challenges in next-generation wireless networks, with a particular focus on cutting-edge areas such as virtual reality (VR), the metaverse, mobile edge computing (MEC), and AI-powered networks. We design system frameworks and leverage advanced design techniques like approximation algorithms, competitive algorithms, and AI algorithms to develop innovative solutions.
Virtual Reality and Metaverse Networks: We focus on optimizing multimedia network transmission efficiency and ensuring immersive user experiences. This includes designing efficient wired and wireless network resource allocation and scheduling schemes, selecting optimal 3D multi-view video transmission and synthesis view, and integrating social networks to determine synthesis parameters and user routing plans.
Mobile Edge Computing: We combine digital twin technology with distributed AI training frameworks to design advanced algorithms for building efficient and reliable social IoT and crowdsourcing systems. We rigorously validate system performance using real-world datasets and AI models, ensuring stability and reliability in practical applications.
AI Network Optimization: In various AI training frameworks like federated learning and graph neural networks, we design dynamic routing, data source selection, training feature selection, and topology control strategies. Our goal is to minimize total bandwidth and computational resource consumption while satisfying network link/node capacity constraints and diverse application requirements.
5G/6G Wireless Network Optimization: Considering mainstream 5G/6G communication technologies such as NR, NOMA, and RSMA, we design resource block allocation algorithms, multicast grouping and selection schemes, and wireless charging scheduling and power allocation strategies, all while taking into account the specific system limitations of each technology.
本實驗室致力於解決次世代無線網路中各類應用問題，深入探索虛擬實境、元宇宙、行動邊緣運算和AI網路等前瞻領域。我們設計系統框架並運用近似演算法、競爭演算法、AI演算法等高階設計技巧，打造創新解決方案。
虛擬實境和元宇宙網路：如規劃有線及無線網路資源配置和排程方式、選定最佳3D多視角影片傳輸及合成之視角、結合社群網路決定合成參數和使用者路由規畫，以最佳化多媒體網路傳輸效率及確保使用者的沉浸體驗。
行動邊緣運算：如將數位雙生 (Digital Twins) 技術與分散式AI訓練架構相結合，設計高階演算法，建構高效、可靠的社群物聯網和群眾外包系統，並用真實資料集和AI模型，嚴謹驗證系統效能，確保其在實際應用中的穩定性和可靠性。
AI網路最佳化：如在聯盟式學習、圖神經網路等不同AI訓練框架下，我們設計動態路由、資料源選擇、訓練特徵選擇和拓樸控制等策略，最小化總頻寬和計算資源消耗，同時滿足網路鍊接／節點容量限制及不同應用需求。
5G/6G無線網路最佳化：如在考量NR、NOMA、RSMA等5G/6G的主流通訊技術下，考量其系統限制，設計資源區塊分配演算法、群播分群和選擇、無線充電排程和能量分配。