Dynamic resource allocation for multimedia communications over distributed systems

点击翻译按钮获取中文摘要

The advances of Internet and wireless access technologies have opened up new opportunities to develop high-quality multimedia applications. Many multimedia applications involve real-time multimedia communications over distributed systems, in which there is no centralized controller. Examples of the distributed systems are Peer-to-Peer (P2P) networks and multi-hop wireless networks. A distributed system is typically dynamic, since the source model, the channel condition, and the overlay topology are time-varying. It is a great challenge to optimize the resource allocation for multimedia communications over a dynamic distributed system. In this project, a new framework for dynamic resource allocation is proposed to address this challenge. Under the proposed framework, each node proactively collects the information about the system, adaptively learns the system models, and then optimizes the actions in a collaborative and distributed manner. The proposed framework will be applied to solve the following six problems: 1) information collection and adaptive learning in a distributed system, 2) achievable video quality in P2P streaming systems, 3) cross-channel resource optimization for multi-channel P2P streaming systems, 4) cross-layer optimization for video streaming over multi-hop cognitive radio networks, 5) video quality maximization for wireless visual sensor networks under power constraints, and 6) context-aware health monitoring by integrating body sensor networks and wireless visual sensor networks.