信号检测理论在无线传感器网络目标探测中的应用研究

发布时间：2018-08-04 15:24

【摘要】：无线传感器网络(WSN,Wireless Sensor Networks)的提出,是基于环境监测和目标追踪等一系列的实际应用需求。作为一种新兴的对环境和目标进行感知与探测的技术手段,如何提高WSN对目标的检测效率是一个十分重要的研究内容。因此,WSN的信号检测问题是其应用和研究的关键问题之一。本文利用信号检测理论的经典方法,以能量检测器(ED,Energy Detector)和协方差检测器(Covariance Detector)为主要研究目标,结合WSN的信号传播模型以及空间相关性理论,采用本文定义的覆盖面积作为性能指标,分析和比较了两种检测器的检测性能。根据构成检测器的WSN节点个数分类,分别从单点独立检测、两点联合检测和三点联合检测三种配置方式,对ED和CD的检测性能进行了理论分析与仿真测试。在单点独立检测和两点联合检测的研究中,理论分析与仿真测试结果相一致。对于ED,使用两点联合检测可以获得比单点独立检测更好的检测性能(更大的覆盖面积)。对于ED和CD进行两点联合检测的性能研究中,ED采用AND和OR两种判决规则。理论分析和仿真结果表明,当节点间的距离较小时,使用CD可以获得比ED更好的检测效果。但是当节点间的距离增大,使用ED的OR方式能获得更好的检测效果。对ED采用AND和OR两种规则进行两点联合检测所获得的覆盖面积进行分析可以得出结论,当节点间的距离小于某一特定值,使用AND规则可以获得比使用OR规则更好的检测性能;反之,当节点间的距离超过这一特定值,使用OR规则可以获得比AND规则更好的检测性能。对于采用ED进行两点联合检测时AND规则与OR规则这一特定的性能分界点,本文从理论上进行了求解与验证。论文还从理论上研究了三点联合检测问题,首先基于ED和CD这两种基本检测器,设计了适用于三点联合检测的检测器。其次为了将三点联合检测的性能与单点独立检测以及两点联合检测的性能进行比较,设计了一种对检测器的目标区域的检测概率分布进行统计与表示的算法。最后通过这种方法,对检测器的单点独立检测、两点联合检测以及三点联合检测的性能进行分析,研究结果表明,使用三点联合检测的检测器性能明显下降。因此,使用两点联合检测可以获得最好的检测效果。
[Abstract]:Wireless Sensor Network (WSNN) Wireless Sensor Networks) is based on a series of practical applications such as environmental monitoring and target tracking. As a new technique to perceive and detect the environment and target, how to improve the detection efficiency of WSN is a very important research content. Therefore, the signal detection problem of WSN is one of the key problems in its application and research. In this paper, the classical method of signal detection theory is used, the energy Detector and covariance detector (Covariance Detector) are the main research targets, and the signal propagation model of WSN and the spatial correlation theory are combined. The detection performance of the two detectors is analyzed and compared by using the coverage area defined in this paper as the performance index. According to the number of WSN nodes which constitute the detector, the detection performance of Ed and CD is analyzed theoretically and simulated from three configurations: single point independent detection, two point joint detection and three point joint detection. In the research of single point independent detection and two-point joint detection, the theoretical analysis is consistent with the simulation results. For EDs, using two-point joint detection can achieve better detection performance than single-point independent detection (larger coverage area). In the performance study of two points joint detection of Ed and CD, AND and OR are used in Ed. Theoretical analysis and simulation results show that CD can obtain better detection results than Ed when the distance between nodes is small. But when the distance between nodes increases, the OR method of Ed can obtain better detection effect. When the distance between nodes is less than a certain value, the detection performance of AND rule is better than that of OR rule. On the other hand, when the distance between nodes exceeds this specific value, the detection performance of OR rule is better than that of AND rule. In this paper, we solve and verify the AND rule and OR rule, which is a specific performance boundary point, when Ed is used for two points joint detection. The paper also studies the problem of three point joint detection theoretically. Firstly, based on Ed and CD two basic detectors, a detector suitable for three point joint detection is designed. Secondly, in order to compare the performance of three-point joint detection with that of single-point independent detection and two-point joint detection, a statistical and representation algorithm for the detection probability distribution of the target region of the detector is designed. Finally, the performance of single point independent detection, two point joint detection and three point joint detection is analyzed by this method. The results show that the performance of the detector using three point joint detection is obviously decreased. Therefore, the best detection effect can be obtained by using two-point joint detection.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5;TP212.9

【参考文献】