| 摘要: | 本整合型計晝將發展智慧城市内環境輔助生活之下世代行動通訊技術，基於智慧城市的願景，我 們預計實現環境輔助生活的安全防護，利用城市中的影像資料，進行智能分析、萃取資訊、以及特定 目標的追蹤與事件偵測，另外搭配佈建於城市中的小細胞基站，透過新式無線通訊技術完成緊急事件 的回報、處理以及串流影像的傳送。為了提供可靠與即時的通訊服務品質，將開發軟體定義網路、低 延遲傳輸技術與巨量多輸入輸出天線陣列等次世代通訊系統中關鍵的技術環節，以實現從應用層、媒 體存取層與實體層間的資訊處理以及訊號傳輸，針對此智慧城市内的環境輔助生活應用提供前瞻的解 決方案並優化使用者經驗。 而本子計晝針對巨量多輸入輸出下行前編碼技術進行研究，預計採用混和前編碼架構，包含只有 相位的射頻端類比前編碼與數位前編碼，來有效降低系統的複雜度，並建立毫米波全方位通道模型以 供系統性能評估使用。數位下行前編碼單元必須能去除多重使用者干擾與達到三度空間波束成型的效 果，並且加入波束追蹤功能以支援行動通訊服務。由於下世代行動通訊系統具高傳輸率與高容量的需 求，最後將設計硬體加速器來增進運算效能與產出率，以達到提升頻譜效能、傳輸效能、與通訊品質 的目標。 ;To achieve the goal of the smart city, information and communication technologies are used to enhance the interactivity of urban services and to reduce the resource consumption. This integrated project will develop the next-generation wireless mobile communication techniques and video processing techniques for ambient-assisted living in the smart city. We aim to build up the safety and security shield for ambient assisted living by taking advantage of the image/video from surveillance cameras near the streets and the dashboard cameras equipped in the cars, and then process the image/video by intelligent analysis, information extraction, multiple target tracking and event detection. Through the dense deployment of small cells in the city, the advanced wireless communication techniques are also studied and adopted to aid the emergence alert, processing procedure, and video streaming. To provide reliable and real-time communication quality, the following key features in the next-generation mobile communication systems will be developed, including software-defined network, low end-to-end link latency, and massive multiple-input multiple-output antenna array. The signal flow of the integrated system covers from application layer, media-access-control layer, and physical layer. The solution for ambient-assisted living in smart cities will be provided to enhance user experience. The sub-project aims to develop the massive multiple-input multiple-output downlink precoding technique, which refers to a large number of antennas equipped in the base-station. To reduce the system complexity, hybrid precoding architecture will be adopted, which uses analog precoding by phase shifters in the radio-frequency front-end and digital precoding in the baseband. The full-dimension milli-meter wave channel model, including the distribution statistics of power, delay, departure/arrival azimuth angle, and zenith angle, will be constructed for system performance evaluation. The precoding unit at the transmitter side of the base-station must remove or suppress the multiple access interference and also the inter-cell interference. Besides, spatial-division multiple access can be supported with the beamforming effect. The time-varying channel response due to mobility will be also considered and thus beam tracking algorithm will be studied and designed. Because high-throughput is the targets of the next-generation mobile communication systems, the precoding scheme will be implemented as a hardware accelerator in 40nm CMOS technology to demonstrate the feasibility. Hence, the design striking a good balance between complexity and performance is the major concern. The use of massive MIMO in wireless network has the potential to improve energy efficiency, spectral efficiency, and link quality. |
