부산대학교 도서관

The proliferation of real-time networked control applications presents significant challenges to current wireless communication networks, owing to the demanding real-time requirements of machine-type communications. To ensure reliable data transmission, error-correction techniques are used to combat the channel error by adding channel coding redundancy, which exacerbates the timeliness of data transmission. This paper takes the lead to investigate task-oriented polar coded hybrid automatic repeat request with incremental redundancy (HARQ-IR) framework for real-time applications. We exploit the spatio-temporal correlation at information source and the inherent redundancy of code domain to quantify the amount of useful information (regarding the real-time status at source) that each HARQ packet brings to the receiver, which is termed as the real-time gain of transmitting a coded status update. Based on the theoretical analysis, we propose a real-time gain based blocklength optimization scheme to strike a balance between timeliness and reliability by dynamically adjusting blocklength based on channel conditions and dynamics of observed process. Simulation results demonstrate that the proposed scheme outperforms other typical error control schemes in terms of the real-time estimation accuracy, especially in the low signal-to-noise ratio (SNR) region. This paper paves the way to the design of timely error control techniques for machine-type communications.