| （英） |
In recent years, the increasing expansion of internet content and the proliferation of IoT devices have led to a sharp rise in communication demand. In particular, the need for efficient transmission of rich content, such as images and videos, has been rapidly growing.
To address this issue, DeepJSCC (Deep Joint Source-Channel Coding), which integrates source coding and forward error correction using Deep Learning, has garnered significant attention. The rapid development of Deep Learning is expected to improve the transmission quality of video content.
This study adapts DeepJSCC to MIMO environments and evaluates its robustness under mobile conditions. In the training phase, training is conducted in the stationary condition and transmissions as the testing phase are performed in time-varying channels.
This presentation reports the preliminary results of evaluating the impact of terminal speed and the number of receiving antennas on the quality of transmission, particularly in scenarios where the receiving terminal is in motion. |