ITR: Collaborative Research: Optimization and Integrated Control of Low Power Wireless Multimedia Networks

近年来，多媒体设备和通信技术的数量出现了爆炸性的增长。虽然这些多媒体技术仍然依赖于有线链路，但多媒体通信的未来是无线的。移动多媒体网络将提供漫游用户之间的无缝通信。视频会议或互联网接入等应用程序将在世界上任何地方都可以使用。多媒体应用程序将生成随时间变化的大不同类别的网络流量(流量异质性)。网络本身将由有线和无线链路组成(网络异构性)。我们已经开始感受到这种变化。多媒体无线技术的一个主要限制因素是有限的功率预算。笔记本电脑和个人数字助理等移动设备的电池寿命有限。可充电电池技术只带来了电池效率的小幅提高。因此，为了保证基于无线的多媒体通信的持续增长，优化网络协议以达到最低功耗是非常重要的。为了实现这一目标，已经在网络协议栈的各个层进行了许多尝试。这些努力大多仅集中在物理层和链路层。在该方案中，我们将无线网络协议栈作为一个整体来处理，并提供了一个统一的框架来优化协议栈各个层的计算和传输功率。电源管理器是我们框架的核心。协议栈的不同层通过电源管理器相互通信，电源管理器根据网络、流量和电源限制做出自适应策略决策。这是该提案的一个独到之处。在传统模型中，网络层只有很少的通信或根本没有通信，导致与所提出的框架相比性能较差。所提出的方法致力于在移动无线节点处以最小的功耗来优化服务质量(QoS)。为了实现我们的目标，我们考虑了以下研究问题：物理层的自适应信源编码和调制策略，以最小化功率以获得所需的可靠性。自适应是基于链路层产生的信道状态估计来完成的。新的容错编码方法不具有传统前向纠错码的缺点。低开销、优雅的降级和简单的编码和解码降低了计算的功率需求，是该编码器的主要优势。这些编码器特别适合最先进的压缩标准，如JPEG、mpeg和H.263。还提出了完全不同的在线信道状态估计器。区分不同信道和电池状态的功率感知传输控制协议。应用层自适应机制来节省功率。这些机制使用网络和信号处理领域的技术来提高性能。自适应电源管理器(集成控制器)策略，从全局角度控制整个无线网络。策略决策考虑了可靠性、带宽、健壮性和功耗。这是一个模拟框架，用于试验不同的硬件配置以及网络拓扑和约束。所开发的框架将有助于验证和更新我们的理论。这一将通过网络提供的工具将对其他研究人员设计未来节能的无线多媒体系统具有无价的价值。

Recent years have seen an explosive growth in the number of multimedia devices and communication tech-nologies. While these multimedia technologies still depend on wireline links, the future of multimedia com-munication is in wireless. Mobile multimedia networks will provide seamless communication between roam-ing users. Applications such as video conferencing or internet access will become available from any where in the world. Widely varying classes of network traffic (traffic heterogeneity) that change over time will be generated by multimedia applications. The network itself will consist of both wireline and wireless links (network heterogeneity). We are already beginning to feel this change.A major constraint in multimedia wireless technology is the limited power budget. Mobile devices such as laptops and personal digital assistants have limited battery life. The re-chargeable battery technology has resulted only in a small increase in battery efficiency. So, it is important to optimize network protocols for minimum power to ensure the continued growth of wireless based multimedia communication. Many attempts have been made towards this goal at individual layers of the network protocol stack. Most of these efforts have concentrated only at the physical and link layers. In this proposal, we deal with the wireless network protocol stack as a whole and provide a unified framework to optimize both computation and transmission power at various layers of the protocol stack. A power manager serves as the central core of our framework. Thedifferent layers of the protocol stack communicate with each other through the power manager that makes adaptive policy decisions based on the network, traffic, and power limitations. This is a uniqueness of the proposal. In the conventional model, the network layers have minimal or no communication at all leading to poor performance compared to the proposed framework. The proposed methods strive to optimize the quality of service (QoS) with minimum power consumption at the mobile wireless nodes. Towards achieving our objectives we consider the following research issues :Adaptive source coding and modulation strategies at the physical layer that minimize power for a desired reliability. The adaptation is done based on channel state estimates produced by the link layer.New error-resilient coding methods that do not have the drawbacks of the conventional forward error correcting codes. Little overhead, graceful degradation, and simple encoding and decoding that reduce the power requirements for computations are the main strengths of the coder. The coders are especially well-suited for state-of-the-art compression standards like JPEG, MPEG, and H.263. Radically different on-line channel state estimators are also proposed.Power-aware transport control protocols that distinguish between various channel and battery state.Application layer adaptation mechanisms to conserve power. The mechanisms use techniques from networking and signal processing areas to improve the performance.Adaptive power manager (integrated controller) policies that control the entire wireless network from a global point of view. Reliability, bandwidth, robustness and power considerations are taken into account for policy decisions.A simulation framework to experiment with different hardware configurations and network topologies and constraints. The developed framework will be help to validate and update our theories. This tool that will be made available through the web will be invaluable for other researchers in designing future energy-efficient wireless multimedia systems.