CAREER: M-ary Distributed Detection in Wireless Sensor Networks

本研究的重点是集成多尺度分布式检测和通信，对无线传感器网络（wsn）和认知无线网络（crn）的理论和实践具有重要影响。重点解决了以下四个问题：(1) M-ary分布式检测：PI针对不同无线信道模型的M-ary分布式检测系统开发融合规则和局部决策规则，研究错误信道如何从根本上限制这些系统的性能。在此基础上，研究了应用于wsn和crn的复合多源分布式检测问题：多声源联合检测与定位以及多通道系统中的频谱感知。(2)无线分布式检测系统的顺序方法：为了深入理解无线网络中静态和顺序分布式检测的权衡，本课题设计了两种仅在传感器处进行顺序测试和仅在融合中心进行顺序测试的顺序分布式检测系统，并研究了所采用的无线信道模型对静态和顺序检测系统的检测可靠性和延迟的影响。(3)不完全信道状态信息对无线分布式检测系统性能的影响：PI研究了信道估计误差对静态和顺序分布式检测系统设计的影响，并研究了这些误差如何进一步限制系统检测性能。(4)无线分布式检测系统的协作：针对静态分布式检测系统和顺序分布式检测系统，提出了基于相邻节点间局部信息传递的协作方案，探讨了协作在检测可靠性和时延方面提供增益的条件，并寻求了协作增益与不同信道模型之间的关系。

This research focuses on integrating M-ary distributed detection and communications, and makes a significant impact on the theory and practice of wireless sensor networks (WSNs) and cognitive radio networks (CRNs). In particular, the following four problems are tackled.(1) M-ary distributed detection: The PI develops fusion rules and local decision rules for M-ary distributed detection systems with different wireless channel models, and studies how the erroneous channels fundamentally limit the performance of these systems. Furthermore, two composite M-ary distributed detection problems that have applications in WSNs and CRNs are investigated: joint detection and localization of multiple acoustic sources and spectrum sensing in a multi-channel system. (2) Sequential methods for wireless distributed detection systems: Striving to gain a profound understanding of the tradeoffs between static and sequential distributed detection in wireless networks, the PI designs two sequential distributed detection systems with i) sequential tests only at the sensors and ii) sequential test only at the fusion center, and studies how the detection reliability and delay of static and sequential detection systems are affected by the adopted wireless channel model. (3) Impact of imperfect channel state information on the performance of wireless distributed detection systems: The PI studies the effects of channel estimation errors on the designs of static and sequential distributed detection systems and investigates how these errors further limit the system detection performance. (4) Cooperation in wireless distributed detection systems: For both static and sequential distributed detection systems cooperative schemes are developed, based on local information passing among neighboring nodes, and conditions under which cooperation provides gain, in terms of both detection reliability and delay, are explored and the relationships between the cooperative gain and different channel models are sought.