CAREER: Energy Efficient Mobile Wireless Networking

移动无线网络正以惊人的速度增长。许多传统的有线应用正在走向无线化。在其他问题中，未来移动无线通信成功需要解决的一个核心问题是节能无线网络。目前，由于各种原因，大多数移动设备的可充电电池能量容量受到限制。众所周知，与无线技术和应用的进步相比，电池技术的进步速度相当缓慢。因此，优化网络算法、协议和参数对提高网络能效具有重要意义。特别是，这对于支持下一代高速无线网络可能是必不可少的。能量最小化有两种类型：通信能量最小化和计算能量最小化。在设计节能的物理层算法和硬件架构方面已经取得了重大进展。但是，这个项目的目标是研究网络协议栈更高层的能量优化和控制策略。为了实现这一目标，我们提出了一个叫做控制层的新层。这一层采用了丰富的数学工具和随机控制算法来预测、估计和控制时变参数，如电池容量、传输速率、错误控制、路由、传输协议和应用需求。该层使其他协议层能够感知能量。综上所述，该跨学科研究项目的建议计划以及将研究与本科和研究生水平的教学相结合的教育举措如下：1)时变无线链路的节能链路和MAC层协议。通信与计算能量之间权衡的随机控制。新的基于概率的MAC协议，尝试调度用户，使能量最小化。2)针对突发交通和链路错误的随机控制算法的理论分析。实验验证和测试这些控制算法在现实生活中的频道。3)考虑移动节点当前电池状态和网络状态的随机最小能量路由算法。4)能量感知传输控制协议，可以区分无线链路的好状态和坏状态，并采取相应的行动。5)基于随机动态规划的应用层速率控制技术，实现节能。6)在课堂上实施项目化学习和主动学习技术。通过NSF CAREER项目激励本科生和研究生进行研究。史蒂文斯研究所和新泽西州已分别为该项目提供了10万美元和10万美元的配套赠款。已附上证明文件作为补充。

Mobile wireless networking is experiencing a tremendous rate of growth. Many traditional wireline based applications are going wireless. Among other issues, one core issue that needs to be addressed for the success of future mobile wireless communications is energy-efficient wireless networking. Currently, the re-chargeable battery energy capacity of most mobile devices is limited due to a variety of reasons. It is also known that the rate at which battery technology isprogressing is rather slow compared to the advances in wireless technology and applications. Therefore, it is important to optimize the networking algorithms, protocols and parameters for energy efficiency. Especially, this may be essential to support next generation high speed wireless networking.Energy minimization is of two types: communication and computation energy minimization. There has been significant progress in designing energy-efficient physical layer algorithms and hardware architectures. But, the goal of this project is to investigate energy optimization and control policies at higher layers of the network protocol stack. Towards this objective a new layer called Control Layer is proposed. A rich set of mathematical tools and stochastic control algorithms are incorporated in this layer to predict, estimate, and control the time-varying parameters such as battery capacity, rate of transmission, error control, routing, transport protocols, and application requirements. This layer enables the other protocol layers to be energy-aware.In summary, the proposed plan for this inter-disciplinary research project and the educational initiatives to integrate research with teaching at both undergraduate and graduate levels are as follows:1) Energy-efficient link and MAC layer protocols for time-varying wireless links. Stochastic control of the tradeoff between communication versus computation energy. New probability based MAC protocols that attempt to schedule users such that energy is minimized.2) Theoretical analysis of stochastic control algorithms that react to the bursty nature of traffic and link errors. Experiments to validate and test these control algorithms for real-life channels.3) Stochastic minimum energy routing algorithms that account for the current battery state of a mobile node and also the network state.4) Energy-aware transport control protocols that distinguish between the good and bad states of a wireless link and act accordingly.5) Stochastic dynamic programming based rate control techniques at the application layer to save energy.6) Implementation of project based learning and active learning techniques in the class room. Motivate undergraduate and graduate student research through this NSF CAREER project.A matching grant of $100,000 from Stevens Institute and $100,000 from New Jersey state have been committed to this project. Supporting documents have been attached as supplements.