CAREER: Cross-Layer Architectures for Power Adaptive and Smart Antenna Equipped Mobile Ad Hoc Networks

该方案的目标是通过有效地使用功率控制和智能天线来增加自组织网络的容量。传统上，协议栈各层之间有明确的分离。特别是，物理层和更高层协议的设计和开发往往是独立进行的。为了利用特定层提供的功能，让其他层了解其他层的特性是很重要的。特别是，传统协议不能有效地利用调谐传输功率或智能天线的存在的能力。事实上，它们中的许多都不能正常工作或性能下降。PI提出了跨层架构，该架构将由MAC、路由和传输层上新的相互依赖的协议组成，这些协议可以有效地利用上述物理层的特性。首先，提出了一种功率感知的跨层架构，其中各层共同协作，根据短期和长期因素动态选择正确的传输功率水平。短期电源管理的基础是暂时提高功率水平，以重新建立由于移动而失败的链路，从而为较高层协议提供短暂的宽限期，以便从故障中恢复过来。为此，提出了有助于确定链路故障是由于移动还是由于拥塞效应引起的方法。基于这些决定，就应该如何调整功率水平做出明智的决定。长期电源管理基于选择长时间的传输功率水平，从而形成网络拓扑以实现最佳吞吐量。功率级别的选择是基于各层之间的交互共同做出的，这样就可以在吞吐量和功耗等性能指标之间实现适当的权衡。我们的方法适用于网络中节点在功率能力方面是异构的情况，即不同节点能够传输的最大功率可能不同。其次，提出了一种利用智能天线存在的跨层架构。该体系结构促进了MAC层、路由层和传输层之间的密切交互，以在高度动态的环境中利用定向传输和接收。这些交互有助于选择正确的本地连接，以确保实现频谱的高水平空间重用，同时保持通信开销相当小。该教育计划旨在提高学生对无线网络的认识，并激励他们进一步探索无线网络这一令人兴奋的领域。此外，预计外展活动将提高当地社区对无线网络的认识。更广泛的影响：该研究将通过弥合物理层和更高层技术之间的差距，对自组织网络设计产生重大影响。预计我们将开发的跨层架构将显著增强自组织网络的能力。这反过来又有望进一步扩大这些网络广泛部署的潜力。这种部署的例子可能是车辆网络、灾难恢复任务和战术战场任务。此外，预计将开发的机制将提供相当大的电力节省，从而减少电池消耗，从而延长网络的使用寿命。预计这项工作将激发其他可能的跨层技术方面的未来研究和开发，其中更高层协议可以利用其他物理层功能，如可变调制技术或动态改变错误控制编码级别的能力。研究议程由一个强有力的教育计划补充，包括：(a)加强现有网络课程的内容，并在本科和研究生阶段开设新的、令人兴奋的无线网络课程；(b)增加我们的外展活动，通过在当地高中的演示和讲座，将无线技术带到当地社区；(c)与工业界密切合作，提高学生对商业系统的认识。主要智力贡献：本提案的主要智力贡献是利用跨层依赖来提供层之间的交互，以充分利用移动自组织网络中物理层的特征。该提议者特别关注智能天线的存在以及在物理层调整传输功率水平的能力。如果不提供这种交互，并且不同层的协议是相互独立设计的，那么网络的容量将大大低于这种交互所能实现的容量。此外，许多协议可能无法按照设想的方式工作，或者可能无法令人满意地执行。我们提供了一种机制，允许各层合作，智能地做出选择(a)网络拓扑应该如何开始形成和动态调整，(b)在特定层上做什么来响应另一层的事件，以增加网络的容量并实现相当大的节能。

In this proposal the goal is to increase the capacity of ad hoc networks by efficiently using power control and smart antennas. Traditionally there has been a clear separation between the layers of the protocol stack. In particular, the design and development of protocols at the physical and higher layers has tended to proceed independently. In order to exploit the functionalities provided by a particular layer it is important that the other layers be made aware of the features at the other layers. In particular, traditional protocols do not exploit the ability to tune transmission power or the presence of smart antennas effectively. In fact, many of them fail to function correctly or suffer from degraded performance. The PI propose cross-layer architectures that would consist of new inter-dependent protocols at the MAC, routing and the transport layers that can effectively exploit the aforementioned physical layer features. First the proposer propose a power aware cross-layer architecture wherein the layers jointly co-operate to dynamically choose the correct transmission power levels in response to both short term and long term factors. Short-term power management is based on increasing the power levels temporarily to re-establish links that fail due to mobility to provide the higher layer protocols a short grace period to recuperate from failures. Towards this the proposer propose methods that help in determining whether link failures are due to mobility or are induced due to congestion effects. Based on such determinations, intelligent decisions on how the power levels ought to be tuned are taken. Long-term power management is based on choosing transmission power levels for long periods of time such that the topology of the network is shaped to achieve the best throughput. The choice of the power levels are jointly made based on interactions between the layers such that appropriate trade-offs between performance metrics such as throughput and power consumption are achieved. Our methods are applicable in cases wherein the nodes in the network are heterogeneous in terms of power capabilities, i.e., the maximum powers with which different nodes are able to transmit are potentially different. Second, the proposer propose a cross-layer architecture that exploits the presence of smart antennas. The architecture facilitates close interactions between the MAC, routing and transport layers to exploit both directional transmissions and receptions in a highly dynamic environment. The interactions help choose the right local connectivity to ensure that a high level of spatial re-use of the spectrum is achieved while keeping the communication overheads fairly small. The educational plan is expected to enhance awareness of wireless networking among their students and motivate them to further explore the exciting field of wireless networking. In addition, outreach activities are expected to increase the awareness of wireless networking in the local community. Broader Impact:The research will have a big impact on ad hoc network design by bridging the gap between the physical layer and higher layer technologies. It is expected that the cross layer architectures that we will develop will lead to a significant enhancement in the capacity of ad hoc networks. This in turn, may be expected to further the potential of the widespread deployment of these networks. Examples of such deployments may be in vehicular networks, disaster recovery missions, and tactical battlefields missions. Furthermore, the mechanisms that will be developed are expected to provide considerable power savings that would reduce battery drainage and thereby extend the longevity of the network. It is envisioned that this work will motivate future research and development in terms of other possible cross-layer technologies wherein higher layer protocols can exploit other physical layer functionalities such as variable modulation techniques or the ability to dynamically vary the level of error control coding. Research agenda is complemented by a strong educational plan that includes (a) the enhancement of the content of existing courses in networking and the initiation of new and exciting courses on wireless networks at both the undergraduate and graduate levels (b) increase our outreach activities to bring wireless technology to the local community by demonstrations and talks at local high schools and (c) close collaboration with industry to improve our students' awareness of commercial systems.Major Intellectual Contributions: The major intellectual contribution in this proposal is to exploit cross-layer dependencies to provide interactions between layers to fully exploit features at the physical layer in mobile ad hoc networks. In particular the proposer focus on the presence of smart antennas and the ability to tune transmission power levels at the physical layer. If such interactions are not provided and protocols at the different layers are designed independent of each other, the capacity of the network is considerably lower than what can be achieved with such interactions. Furthermore, many of the protocols may not work in the way envisioned or may perform unsatisfactorily. We provide mechanisms that can allow the various layers to co-operate to intelligently make choices on (a) how the network topology should be formed to begin with and tuned dynamically and, (b) what to do at a particular layer in response to an event at another layer, in order to increase the capacity of the network and achieve considerable power savings.