Two-Timescale Adaptive Live Video Streaming Transmission Mechanism for Vehicular Networks

This research explores a novel adaptive live video streaming transmission strategy over vehicular networks to solve the conundrum between resource-constrained environment and user demand on high Quality-of-Experience (QoE). With an exquisite design of resource types and channel variations, we propose a two-timescale transmission mechanism which allocates the bitrate and bandwidth for each large-timescale frame based on the statistical knowledge of the channel state information (CSI) and refines the power allocation for each small-timescale slot based on instantaneous CSI. Subsequently, we formulate a QoE maximization problem under the restrictions of finite bitrates, bandwidth and power budget. To solve this problem with low computation complexity, we propose a two-stage online successive convex approximation (TOSCA)-based resource allocation algorithm. Simulation results illustrate the rationality and necessity of the proposed dual-time scale optimization, and the proposed mechanism noticeably outperforms the conventional resource allocation schemes under different power budgets and lane configurations.