Integrated Sensing: Network Support for Distributed Sensing Applications

下一代网络应用和技术的核心在于在网络中嵌入传感和执行能力的潜力，从而支持各种监测和控制应用。在此建议中，我们针对需要严格服务质量保证的应用程序，例如，需要物理属性状态的全局快照的应用程序，或需要关联各种此类属性的应用程序。例如，人们可以联合跟踪风速和给定病原体的密度来创建进化模型。为了支持这些应用，同时实现高传感器利用率，我们建议研究两个互补的问题：(1)从网络传感器提取时空数据的估计和控制应用的基本特征和要求是什么；(2)什么类型的网络支持，例如，感知“原语”和资源/流量管理机制对于实现此类网络的广泛共享使用至关重要。由于传感器网络的预期异质性以及传感器处理查询的“能力”与网络带宽延迟产品之间可能存在的不匹配，因此从多个用户共享的潜在最大传感器中实现及时协调响应可能很困难。与此类感知活动相关的通信的固有不对称加剧了这种情况，即一对多查询导致多对一传感器响应。因此，可能会出现不可预测的流量突发、网络热点和性能下降。作为研究的一部分，我们将从应用、网络和传感器的角度来研究这些问题。

At the core of the next generation of networking applications and technologies lies the potential of embedding sensing and actuating capabilities in the network so as to support a variety of monitoring and controls applications. In this proposal we target applications that require stringent quality of service guarantees, e.g., applications needing global snapshot(s) of the state of a physical attribute,or requiring correlating various such attributes. For example, one might jointly track the wind speed and the density of a given pathogen to create evolution models. To support such applications while achieving high sensor utilization we propose to investigate two complementary problems:(1) What are fundamental characteristics and requirements ofestimation and control applicationsdrawing spatio-temporal data from networked sensors; and(2) What types of network support, e.g., sensing `primitives' and resource/traffic management mechanisms will be critical to enable wide-spread shared usage of such networks.Due to the expected heterogeneity of sensor networksand possible mismatch between a sensor's `capacity' to processqueries and the network bandwidth delay product,achieving timely coordinated responses from a potentially largeset of sensors shared by multiple users may be difficult.This is exacerbated by the inherent asymmetry of communications associated with such sensing activities, i.e., one-to-many queries resulting in many-to-one sensor responses. As a result one can expect unpredictable bursts of traffic, network hot spots, and degradation in performance. As part of the proposed researchwe will investigate these problems from the perspective ofthe applications, the network and the sensors.