【作者】 陈维克；

【导师】 李文锋；

【作者基本信息】 武汉理工大学 ， 机械制造及自动化， 2009， 博士

【摘要】 无线传感器网络是新一代的传感器网络,具有非常广泛的应用前景,其发展和应用将会给人类的生活和生产的各个领域带来深远影响。它的出现将会给人类社会带来极大的变革。本论文针对服务于机器人应用的无线传感器网络路由和节点定位技术进行了系统、深入的研究。主要内容有：1、研究了机器人与无线传感器网络组合工作模式下,移动机器人与传感器网络进行信息交互的网络结构,将移动机器人视为agent,在层次化的网络结构下,实现移动机器人对无线传感器网络资源的访问,使机器人在无线传感器网络的协同下完成导航和任务执行。2、分簇路由算法研究。首先提出了基于QoS的分簇路由算法,该算法采用双簇头模型,以保障无线传感器网络的可靠性。并且以双簇头模型为基础,建立了无线传感器网络的负载均衡机制,达到节点间的能量平衡消耗,防止拥塞；其次,构建了一种两阶段成簇的分簇路由算法。采取能量感知和局部信息集中式的簇头选举机制,使得簇头的选举更加灵活与合理。算法还通过平衡簇内节点数实现负载均衡；第三,提出了基于K-means分簇路由算法(DSCA)。在算法中提出了节点的同步失效概念。DSCA通过K-means得到更加平衡的分簇。发展了一种全新的基于信号接收强度的簇质心求解方法。DSCA采用局部信息集中式的动态多簇头选举机制和动态TDMA通讯轮数分配机制,上述机制使无线传感器网络的能量消耗达到了高度均衡,保证了节点的同步失效。3、提出基于RSSI的无线传感器网络加权质心定位算法。通过对无线电传播路径损耗模型的分析,用信标节点对未知节点的不同影响力来确定加权因子,以提高定位精度。并且在理论分析的基础上,设计了优选信标节点进行节点定位计算的规则,以进一步提高节点定位精度。算法具有简单,计算量小、精度较高的特点。4、提出了利用移动信标,将加权最小二乘估计与扩展卡尔曼滤波(EKF)组合,进行未知节点定位的算法(HLA)。HLA首先利用加权最小二乘估计(WLSE),获得无线传感器网络未知节点的初始位置,再用扩展卡尔曼滤波进一步提高定位精度。提出了WLSE的加权因子的确定方法。同时,HLA还分析并构建了移动信标位置参与EKF迭代计算的最优排序方案。更多还原

【Abstract】 Wireless sensors network, which has a great future, is a new generation of sensor networks. Wireless sensor network has a broad spectrum of applications in many areas. It will bring on significant changes in our life. In this paper, we provide some new idea and new method about wireless sensor networks based on the study of authors, the main content of this dissertation is as follows:1. Network architecture of WSN which serves mobile robots has been studied. Hiberarchy is accepted to achieve information access between mobile robots and WSN. Based on the information access, robots implement their tasks such as navigation.2. Research has been done on Clustering routing protocol. Firstly, a routing protocol based on QoS was suggestted, in which the dual cluster-head model was adopted to improve the reliability and the steadiness of wireless sensor networks. The dual cluster-head model helps the routing protocol to balance energy consumption in all nodes and to avoid congestion at cluster-heads by distributing evenly nodes in all clusters:Secondly we developed a two-phase cluster formation algorithm, which was energy-aware. A local-centralized mechanism was suggested to elect cluster-heads. This mechanism made the election of cluster-head more flexible and reasonable. This algorithm also achieved load balance by distributing evenly the nodes in clusters; Thirdly, the Dynamic Schedule Clustering Algorithm (DSCA) was proposed for clustering nodes in wireless sensor networks based on K-means. In the algorithm, the concept of synchronous failure of sensor nodes was present. Balanced clustering is achieved by using the K-means method, which forms the basis of balancing energy consumption. A new method was developed in DSCA to get the centroid of a cluster by Received Signal Strength Indicator information. DSCA suggestted a mechanism of dynamic multi-clusterhead election and a mechanism of dynamic Time Division Multiple Access communication load schedule that are both local-centralized. These mechanisms ensure highly balanced energy consumption and synchronous failure of nodes.3. Weighted Centroid Localization Algorithm Based on RSSI for Wireless Sensor Networks Localization of nodes is supported. By analyzing the model of radio wave propagation loss, this paper suggests a weighted coefficients, which are decided by the influence of beacons to unknown nodes, to prompt localization accuracy. A criterion which is used to select beacon nodes for computing the position of nodes is also suggested in order to improve localization accuracy more. The algorithm was simple and with low computational complexity and good localization accuracy.4. A hybrid localization algorithm (HLA) was proposed based on mobile beacons. The hybrid algorithm combined weighted least squares estimation with Extended Kalman Filter (EKF). Firstly, the algorithm got an inaccurate position coordinate of node by using weighted least squares estimation. Then the algorithm achieved an accurate localization of node by using Extended Kalman Filter. An approach was proposed to determine the weight exponent. And an optimal sorting solution of beacon position was proposed for EKF iteration.更多还原