In this paper, we propose a novel wireless caching scheme to enhance thephysical layer security of video streaming in cellular networks with limitedbackhaul capacity. By proactively sharing video data across a subset of basestations (BSs) through both caching and backhaul loading, secure cooperativejoint transmission of several BSs can be dynamically enabled in accordance withthe cache status, the channel conditions, and the backhaul capacity. Assumingimperfect channel state information (CSI) at the transmitters, we formulate atwo-stage non-convex mixed-integer robust optimization problem for minimizingthe total transmit power while providing quality of service (QoS) andguaranteeing communication secrecy during video delivery, where the caching andthe cooperative transmission policy are optimized in an offline video cachingstage and an online video delivery stage, respectively. Although the formulatedoptimization problem turns out to be NP-hard, low-complexity polynomial-timealgorithms, whose solutions are globally optimal under certain conditions, areproposed for cache training and video delivery control. Caching is shown to bebeneficial as it reduces the data sharing overhead imposed on thecapacity-constrained backhaul links, introduces additional secure degrees offreedom, and enables a power-efficient communication system design. Simulationresults confirm that the proposed caching scheme achieves simultaneously a lowsecrecy outage probability and a high power efficiency. Furthermore, due to theproposed robust optimization, the performance loss caused by imperfect CSIknowledge can be significantly reduced when the cache capacity becomes large.
展开▼
