Nowadays, real-time wireless video transmission has become increasingly popular and important. However, the uncertainty of wireless channel often causes errors, leading to a loss of freshness in the video transmission. The state-of-the-art video transmission systems overcome this problem by transmitting I-frames periodically to ensure freshness. Unfortunately, I-frames occupy more channel resources, for they do not utilize the information from previous frames. On the other hand, another coding mode named P-frame can save channel resources, but it may cause loss in frame freshness. Therefore, there exists a tradeoff between freshness and channel resources. Most traditional optimization methods only consider bitrate and distortion. In this paper, we propose a novel optimization method that takes information freshness into account. The proposed method employs the Rate-Distortion-Freshness (RDF) function which combines traditional Rate-Distortion (RD) function and information freshness. We analyze using the RDF function for video transmission in multi-state erasure channels and adopting Q-learning to achieve online frame mode selection based on the RDF function. Simulations show that for real-time wireless video transmission, our proposed method reduces channel occupancy without decreasing the quality of video.
RL-based Freshness-aware Frame Mode Selection for Real-time Wireless Video Transmission
2023-06-01
1782143 byte
Conference paper
Electronic Resource
English
Real-Time Transmission of H.264 Video over 802.11B-Based Wireless ad hoc Networks
Springer Verlag | 2005
|Content-Aware Packet-Level Interleaving Method for Video Transmission over Wireless Networks
British Library Conference Proceedings | 2005
|Information freshness-aware task offloading in air-ground integrated edge computing systems
BASE | 2022
|Microwave-based wireless real-time transmission system and aircraft
European Patent Office | 2020
|