Stochastic joint rate control and resource allocation for wireless video surveillance
Guanglun Huang,Baoxian Zhang,Zheng Yao,Cheng Li
DOI: https://doi.org/10.1016/j.comnet.2021.107904
IF: 5.493
2021-05-01
Computer Networks
Abstract:<p>Wireless video surveillance has become increasingly prevalent due to the rapid development of cellular communication technology and the low deployment cost of cellular surveillance cameras. However, it is still a challenging task to deliver high-quality low-latency surveillance videos in resource-limited cellular networks due to the fluctuating characteristics of wireless channels. In this paper, we consider a wireless video surveillance system consisting of one base station and multiple cellular surveillance cameras, which upload their videos to the base station. To optimize the long-term system performance in terms of surveillance video quality and latency, we formulate a stochastic optimization problem that jointly considers video rate control and uplink resource allocation. The formulation captures the tradeoff between video quality and delivery latency while considering the different time scales of video rate control and channel resource allocation. With the application of Lyapunov optimization, we transform the problem into a sequence of per-slot deterministic problems, and propose an online joint rate control and resource allocation algorithm to solve these per-slot problems. Specifically, the proposed algorithm decomposes each of the per-slot problems into two independent subproblems, namely, the rate control subproblem at large time scale and the uplink resource allocation subproblem at small time scale. We derive the closed-form solution for the rate control subproblem, and design a fast-convergent iterative algorithm to solve the uplink resource allocation subproblem optimally. We prove the asymptotical optimality of our proposed algorithm. Extensive simulation results validate our theoretical analysis and demonstrate that the proposed algorithm can well balance the surveillance video quality and delivery latency.</p>
computer science, information systems,telecommunications,engineering, electrical & electronic, hardware & architecture