返回
类型 基础研究 预答辩日期 2017-11-19
开始(开题)日期 2013-11-26 论文结束日期 2017-09-07
地点 中国无线谷1319 论文选题来源 其他项目    论文字数 2.6 (万字)
题目 无线传感器网络的定位算法研究
主题词 能效,非视距环境,定位精度,无线传感器网络
摘要 近年来, 无线传感器网络(WSNs)在全球范围内引起了广泛的关注。得益于无线通信技术的提高,使得低成本、短距离、能量受限以及多功能的传感器节点得以实现。WSNs作为一种自组织及无需基础设施的无线网络,网络中的节点实时监控周围物理环境,如温度、声音、振动、压力、运动或污染物并通过与其它节点协作将收集到的数据传递至汇聚节点,该汇聚节点对接收到的数据进行观察和分析。一个汇聚节点或基站相当于用户和网络之间的通信接口。典型的 WSN 包含数百甚至成千上万的传感器节点。传感器节点通过无线信号传输完成他们之间的沟通。无线传感器节点一般包括传感和计算设备、无线收发机以及功率组件。 传感器节点将监测收集到的数据通过多跳通信传输至目的地到达网关节点,最终通过互联网或卫星到达管理节点。在这一传输过程中监测收集到的数据可能被多个中间环节处理。用户可以对传感器节点进行配置和管理,也可以发布监测任务以及收集监测数据。 传感器节点可以根据业务需求随机部署,从而可以应用到许多不同的领域如军事目标跟踪和监测、自然灾害救济、生物医学健康监测、危险环境勘探以及地震感应。而实现这些应用的基础和关键就是节点的位置信息。如果节点找不到自己的位置或者不能准确定位,节点收集到的数据对于 WSN 的应用是没有意义的。然而定位算法的定位精度对障碍物非常敏感,非视距(NLOS)环境下节点定位误差会很大。此外,传感器节点的能量消耗十分关键,因为节点一旦被部署好之后就很难或无法补充能量。 为了降低 NLOS 环境对定位精度的影响,同时减少定位过程中的节点所消耗的能量,本文提出的如下有效的解决方法。 1. 针对 WNSs 的功率受限问题,首先提出一个约束优化算法,考虑在保证定位精度的条件下,即满足由贝叶斯费舍尔信息矩阵(B-FIM)定义的预先条件,如何分配锚节点的传输功率。而在定位精度要求较高的情景下,需要将定位精度和能量消耗这两个重要并冲突的性能指标同时作为优化目标。因此,我们采用 NSGA-II 多目标优化算法来解决这种双重问题。利用载波监听多址/冲突避免(CSMA/CA)技术结合请求发送(RTS)/清除发送(CTS)机制的 IEEE 802.11 通信协议对网络中节点的能量消耗进行建模;另一方面,从到达角(DOA)和接收信号强度(RSS)测量估计方差的数学模型推导出 Cramer-Rao 界限(CRB)作为目标节点定位性能的目标函数。仿真结果表明:第一,相对于功率均匀分布模型,基于 B-FIM 的约束优化算法大大降低了网络能量消耗;第二,NSGA-II 算法能够有效解决节点定位精度与能量消耗的平衡问题。 2. 针对 WNSs 中节点功率受限问题,我们提出一种可以同时分配锚节点的发射功率和睡眠时间的定位博弈。众所周知,对于资源分配问题,博弈论作为一种便捷的建模工具,能够提供一个理想的框架来帮助设计高效、高鲁棒性的分布式算法。本文针对节点在基于混合DOA/RSS测量模型的定位算法的定位过程中所消耗的能量进行研究。为了解决能耗问题,在保证目标定位满足一定精度的前提条件下,本文采用Leader-follower Stackelberg 博弈对锚节点联盟(CANs)中各个锚节点的传输功率和睡眠时间进行有效分配。在博弈中,目标作为 follower, CANs 作为 leader。CANs 先采取行动,从而目标采取相应的行动。同时,我们证明存在唯一的 Stackelberg 均衡(Stackelberg Equilibrium, SE)点,并且这个 SE 均衡不同于标准博弈中的纳什均衡(NE)。在 NE 状态中,在其他玩家保持不动的情况下没有个人玩家可以通过偏离NE状态而得到更好的收益。而 SE 作为 leader 的最佳策略是考虑到 follower 对 leader行动的最好响应策略。 在本算法中,我们把博弈分为两个层级:目标节点扮演 buyer-level 博弈,而 CANs 扮演 seller-level 博弈。每位玩家自带自私属性,各自争取最大化自身收益。目标节点可以被视为一个买家通过购买 CANs 的传输功率,在最少成本条件下尽可能最大化其收益。CANs 可以视为卖方,他可以获得关于买家策略的情报并以睡眠时间作为酬劳。仿真结果表明了所提算法的优越性能。 3. 为了减少 NLOS 误差对定位精度的影响,基于非参数信任传输(NBP)方法建立一种在 NLOS 环境下的定位算法。根据 NLOS 误差的分布概率及分布参数的先验信息量,给出了 3 种不同情况下定位问题的最大后验概率。第 1 种情形为理想化情形,即知道 NLOS 环境下的测量及相应的误差分布参数。第 2 种情形为知道任意 2 个节点之间的连接处于 NLOS 环境下的概率及相应的 NLOS 误差分布参数。第 3 种情形为最差情形,只获得测量误差的信息。将所提算法与基于最大似然-退火法(ML-SA)进行了比较,仿真结果表明:在每种情形下所提算法获得的定位精度都远超过基于ML-SA 的定位算法。因此,在 NLOS 传输环境下,采用 NBP 的定位算法可获得较高的定位精度。
英文题目 Investigations on Localization Algorithms for Wireless Sensor Networks
英文主题词 Energy-efficient, NLOS environments, positioning accuracy, wireless sensor networks.
英文摘要 Wireless sensor networks (WSNs) have gained world-wide attention in recent years. Recent advances in wireless communications and electronics have enabled the development of low-cost, short in communication range, limited-power and multi-functional sensor nodes. WSNs can be defined as a self-configured and infrastructure-less wireless networks, which monitors physical or environmental conditions, such as temperature, sound, vibration, pressure, motion or pollutants and cooperatively pass their data through the network to a main location or sink where the data can be observed and analysed. A sink or base station acts like an interface between users and the network. Typically a WSN contains hundreds to thousands of sensor nodes. The sensor nodes can communicate among themselves using radio signals. A wireless sensor node is equipped with sensing and computing devices, radio transceivers and power components. Sensor nodes monitor the collected data to transmit along to other sensor nodes by hopping. During the process of transmission, monitored data may be handled by multiple nodes to get to gateway node after multihop routing, and finally reach the management node through the internet or satellite. The user can configures and manages the WSN with the manage node, publish monitoring missions and collection of the monitored data. The sensor nodes are deployed in large numbers provide unprecedented opportunities for a huge of applications of the WSNs in many different fields such as military target tracking and surveillance, natural disaster relief, biomedical health monitoring, hazardous environment exploration and seismic sensing. The foundation and key of applications of WSNs is the location information of sensor nodes. If nodes can not locate their position, or they are located with inaccurate position, the data which is collected by nodes is not useful for applications of WSNSs. The positioning accuaracy is proved that it is very sensitive to the availability of line-of-sight (LOS), which implies that non-line-of-sight (NLOS) environments can make a considerable error in node location estimate. In addition, energy consumption of wireless sensor nodes is also extremely important problem, because the energy of nodes is difficult or cannot be supplied after deployment. To reduce the effects of the NLOS environments on the positioning accuracy and saving the energy of the nodes consumed in the positioning process, we propose the following solutions. 1. Aim at the problem of power-limited of nodes in the WSN: we propose a constrained optimization algorithm that allocates the transmission power of the anchors while remaining the positioning accuracy satisfies a pre-specified metric, which is defined by Bayesian Fisher Information Matrix (B-FIM). In the scenarios where a high accuracy in target’s location is required, positioning accuracy and energy consumption these two conflict metrics should be balanced, meaning that they should be improved at the same time. Therefore, the Non-domined Sorting Genetic Algorithm (NSGA-II) is used to solve this dual problem. We model the energy consumption of nodes in the network by using the Carrier-Sense Multiple-Access/Collision-Avoidance (CSMA/CA) technique based IEEE 802.11 protocol in combination with Request-To-Send (RTS)/Clear-To-Send (CTS) mechanism to perform the communication between nodes. Otherwise, based on the mathematical model of Direction-of-Arrival (DOA) estimation error variance as a function of Received-Signal-Strength (RSS) we derive the Cramer-Rao Bound (CRB) of achievable accuracy of the target’s estimate position as the objective function of positioning error. The simulation results show that: (1) the B-FIM constrained based optimization localization alogirhtm can significantly reduce the energy consumption of nodes in compared to scheme of uniformly power allocation; and (2) the NGSA-II based localization algorithm can effectively solves the tradeoff between localization accuracy and energy consumption. 2. Aim at the problem of power-limited of nodes in the WSN: we propose an energy-efficient localization game, which can allocate the transmission power and the sleep times of the anchor nodes simultaneously. As we knew, game theory provides the ideal framework to design the efficient and robust distributed algorithms. It is a natural tool for modeling and addressing our problem. We study the energy saving problem of nodes for the hybrid DOA/RSS localization scheme in WSNs. To solve this problem, both the transmission power and the sleep times of the anchor nodes are allocated while the target is required to be localized with a pre-specified accuracy. The leader-follower Stackelberg game is used to model our problem, in which the target is the follower and the set of the target’s neighbor anchor nodes, named CANs is the leader. In this game, the CANs plays as the leader who take actions first, and the target plays as the follower then takes actions accordingly. We prove that there exists a unique Stackelberg Equilibrim (SE) in this game which is different to the Nash Equilirium (NE) in the standard game. A NE is the status where no individual player can do strictly better by deviating, holding the actions of other players. But the SE is the optimal strategy of the leader, given the best response action of the follower to the leader’s strategy, together with the optimal strategy of the follower corresponding to the optimal strategy of the leader. We divide the game into two hierarchical levels: the target plays a buyer-level game and the anchors play a seller-level game. Each player is selfish and wants to maximize its benefit. The target can be seen as a buyer who buys the transmission power from the anchors to maximize its benefit at the least possible cost. CANs can be viewed as a seller who has knowledge of target’s intelligence and gets the sleeping time as the payment. We prove the existence and uniqueness of the SE by giving closed-form expressions for the SE strategies of both the target and the CANs. The superior performance of proposed algorithm is valided through simulation results. 3. Aim at the problem of inaccuracy positioning in NLOS conditions, a non-parametric belief propagation (NBP)-based localization algorithm in the NLOS environments for WSNs is proposed. According to the amount of prior information known about the probability of a range measurement is in NLOS environment and distribution parameters of the NLOS error distributions, three different cases of the maximum a posterior (MAP) localization problems are introduced. Firstly is the idealized case where an understanding of NLOS condition measurements and distribution parameters of the NLOS errors is perfectly known. The second case deals with the scenario where we do not know which ranges are in the NLOS condition, however, we know the probability that a link of a pair of nodes is in NLOS condition and the corresponding distribution parameters of the NLOS errors. The worst case is the third one because only knowledge about noise measurement power is obtained. The proposed algorithm is compared with the maximum likelihood-simulated annealing (ML-SA)-based localization algorithm. Simulation results demonstrate that the proposed algorithm provides good location accuracy and considerably outperforms the ML-SA-based localization alogorithm for every case. Therefore, in the NLOS environments, the localization algorithm can obtain the location estimate with high accuracy by using the NBP method.
学术讨论
主办单位时间地点报告人报告主题
东南大学 2016年11月24号 中国无线谷 Dr. Hua Qian Towards 5G IoT
东南大学 2017年7月10日 中国无线谷 Prof. Zihuai Lin Distributed Resource allocation and Spectrum Sharing in Future Wireless Cellular Networks
东南大学 2012年05月12号 东南大学四牌楼校区 裴氏莺 电子商务中的公钥密码应用
东南大学 2012年05月16号 东南大学四牌楼校区 裴氏莺 An Indoor Localization Method Based on the PSCL- RSSI Estimation in WSN
东南大学 2012年01月04号 东南大学四牌楼校区 裴氏莺 使用Kalman滤波方法的基于RSSI的室内定位和跟踪
东南大学 2017年07月09号 东南大学四牌楼校区 裴氏莺 无线传感器网络定位技术简介
东南大学 2017年4月25日 中国无线谷 Prof. Julian Cheng Energy Efficient Resource Allocation for Non-Orthogonal Multiple Access (NOMA) Wireless Networks
东南大学 2017年7月18 榴园宾馆逸夫馆报告厅 Prof. Xin Yao Many Objective Optimization
     
学术会议
会议名称时间地点本人报告本人报告题目
PacRim 2017 2017年08月21 加拿大维多利亚大学 Energy Considerations for Node Localization
VTC 2016 Spring 2016年05月15 南京Hilton酒店 An Accurate and Energy-Efficient Localization Algorithm for Wireless Sensor Networks
     
代表作
论文名称
Energy-Efficient Localization Game for Wireless Sensor Networks
An Accurate and Energy-Efficient Localization Algorithm for Wireless Sensor Networks
NBP-based localization algorithm for wireless sensor networks in NLOS environments
 
答辩委员会组成信息
姓名职称导师类别工作单位是否主席备注
蔡跃明 正高 教授 博导 解放军理工大学
束锋 正高 教授 博导 南京理工大学
潘志文 正高 教授 博导 东南大学
赵新胜 正高 教授 博导 东南大学
许威 正高 教授 博导 东南大学
      
答辩秘书信息
姓名职称工作单位备注
吴亮 其他 讲师 东南大学