Minimizing Age of Information via Variable Deep Coding for X-Relay Status Update Networks

This paper investigates the minimization of the Age of Information (AoI) in an X-relay status update network using semantic communication. AoI quantifies the freshness of received updates, which is crucial for applications requiring timely information. While traditional bit-oriented communication has been extensively studied for AoI optimization, it primarily focuses on delivering data at the bit level, overlooking the potential benefits of content-aware approaches. In contrast, semantic communication prioritizes the transmission of meaningful content, facilitating more efficient and timely updates. A key challenge in minimizing AoI lies in balancing information accuracy and transmission efficiency. To address this, we adopt the deep learning-based joint source-channel coding (DeepJSCC) framework, and propose a variable deep coding scheme that supports variable-length symbol transmissions, referred to as DeepJSCC-V. Experimental results demonstrate that the DeepJSCC-V framework minimizes the average AoI by achieving the best tradeoff between fast transmission and high update success probability. Overall, this study highlights the effectiveness of semantic communication in achieving both timeliness and accuracy in X-relay status update networks.