應用人工智慧於遠距離8K虛擬實境影像即時傳輸之物聯網(IoT)技術開發
Development of Long-Distance Multi-Camera 8K VR Real-Time Transmission for IoT Applications Using Artificial Intelligence
計畫團隊成員 Members 王暉教授
Prof. Huei Wang
國立臺灣大學電信所
EE Dept., NTU
黃天偉教授
Prof. Tian-Wei Huang
國立臺灣大學電信所
EE Dept., NTU
林怡成教授
Prof. Yi-Cheng Lin
國立臺灣大學電信所
EE Dept., NTU
盧信嘉副教授
Assoc. Prof.
Hsin-Chia Lu
國立臺灣大學電子所
EE Dept., NTU
袁孝維教授
Prof. Hsiao-Wei Yuan
國立臺灣大學森林環資系
FO Dept., NTU
Category
UVA Application, Millimeter Wave High-Speed Circuit and System
技術亮點
Technical Highlights
本計畫目標為實現適用4096 QAM調變技術的37-40 GHz毫米波收發機系統,應用於中長距離的4K/8K高解析度虛擬實境影像無線即時傳輸。因為VR需要提供40Gbps以上的數據通訊流量,本計畫將使用4096QAM的高階毫米波調變技術來提升頻譜的使用效率。其中37-40 GHz將劃分500個不同的頻道,採用人工智慧的寬頻最佳化技術來完成500個8 K頻道的數位自動校正(Digital Calibration)。此收發機系統可以提供100個以上的8K攝影機即時傳輸的需求(每個8K @ 60fps攝影機需要400 Mbps的資料量)。此系統在短距離300m之內,可提供8K @ 60fps即時影像傳輸(共54 Gbps),至於長距離7 km,也可以提供4 K @ 30fps的高畫質影像。 此整合型研究計畫的成果為系統電路半導體晶片(SoC)與天線封裝完成的收發機系統模組(SiP),並搭載於無人飛行載具(UAV)上,執行野生動物資源監測。
The goal of this project is to implement a 37-40GHz mmWave transceiver (TRx) supporting 4096 QAM modulation. The proposed TRx is utilized for long-distance multi-camera 4K/8K VR real-time image data transmission. 4096 QAM mmWave modulation technology is applied to improve spectrum efficiency since the data rate of the VR system costs more than 40Gbps. The 37-40GHz band would be divided into 500 channels, and the AI technology is used for the TRx to carry out digital calibration and optimization. The TRx system can provide real-time image transmission for more than 100 8K cameras. For short distance (less than 300m), the TRx can support 8K @ 60fps real-time transmission; even for long distance (about 7km), it also can support 4K @ 30fps real-time transmission. The achievements of this project is an integrated TRx system module (SoC and SiP). This module would be installed on the UAV and applied for wildlife resources monitoring.