Joint Lossy Compression and Power Allocation for Delay-Sensitive Wireless Communications

Shaoling Hu,Wei Chen

DOI: https://doi.org/10.1109/icc40277.2020.9149165

2020-01-01

Abstract:Delay-sensitive communication has attracted much recent attention in emerging 5G systems that are expected to provide Ultra-Reliable Low-Latency Communications (URLLC). However, it remains open on how to balance the queuing delay and the compression distortion in power-constrained transmission over time-varying channels. In this paper, we present a cross-layer approach to strike the optimal delay-power-distortion tradeoff. More specifically, the compression ratio in the application layer and the data rate in the physical layer jointly adapt to the buffer and channel state. To this end, we formulate a non-linear optimization problem with a constrained Markov Decision Process (CMDP) to minimize the average queuing delay while keeping the distortion and power consumption low. Converting this problem into Linear Programming (LP) gives us both the optimal scheduling parameters and its delay-power-distortion tradeoff. We next show that the tradeoff between delay, power, and distortion could be further improved by introducing the weighted-average cost of power and distortion for each action. The weighted-average costs can be minimized and optimized to be a convex function of the transmission rate. With this convex cost-rate function, we find the optimal threshold-based policy.