无线信道在网络仿真平台中的设计与实现

发布时间：2018-03-10 10:17

本文选题：COOJA　切入点：Contiki　出处：《电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：无线传感器网络由随机分布在监测区域的节点组成,可应用于环境监测、军事、工业控制等领域。网络中节点规模和复杂度随着无线传感网监测应用需求而增大,对大规模无线传感网络仿真工具的要求越来越高。复杂的无线信道使得节点间的通信链路存在随机性和不稳定性,直接对网络仿真评估造成影响。以往无线传感网研究中,侧重于协议和能耗,对信道设计大都基于理想信道模型。COOJA(Contiki OS Java)作为近年来新推出的无线传感网仿真平台,可以高效的仿真上百节点的无线传感网络,但由于其信道模型并不能很好的反映真实环境中无线信道衰落特性,研究开发适用无线传感网应用环境的信道模型显得尤为重要。本文首先介绍了无线信道模型的研究进展和应用情况,分析COOJA中信道模型存在的不足,提出开发集成新的信道模型的可能性。研究了目前已有的基于实测数据统计规律建立的经典信道模型,分析它们各自适用的场景特点。本文分别从大尺度衰落和小尺度衰落两方面,研究设计适用于无线传感网应用环境的信道模型。小尺度衰落中,当信道存在视距传播时接收信号服从Rice分布,结合Clarke模型和Jakes的简化思想构建Rice信道模型,仿真验证了该思想具有服从莱斯分布的统计特性、良好的平稳性和各态历经性且具有计算量小的特点。大尺度衰落基于对数路径损耗,采用Plane-Earth模型,通过与实测双斜率模型在不同环境中的对比,给出修正,并分析了它的适用条件。最终分别设计了适应五种不同环境的信道:平坦水泥地、平坦沙地、平坦草地、低灌木丛、森林环境,但由于后两种环境中会受视距分量存在与否的影响,Plane-Earth模型不能统一描述,因而后两种环境使用单斜率模型。最后,在深入研究大无规模仿真平台COOJA的架构和仿真机制的基础上,使用Java语言设计并实现五种信道模型,完成信号强度预测、误包率计算等功能,为物理层提供信号强度指示和数据包接收判断。将信道模型集成在COOJA仿真平台中与原有信道对比,在预测值的比较上,原有信道预测模型预测值存在较大的差异;在仿真性能方面,通过建立500节点的无线传感网仿真场景,由于新的信道模型计算复杂度更高,总体仿真效率略有降低,但信道模型预测更具有合理性,可用于未来COOJA无线传感网络仿真的信道模型选择之一。
[Abstract]:Wireless sensor network is composed of nodes randomly distributed in monitoring area, which can be used in environmental monitoring, military, industrial control and other fields. The size and complexity of nodes in the network increase with the demand of wireless sensor network monitoring application. The requirements of simulation tools for large-scale wireless sensor networks are becoming higher and higher. The complex wireless channels make the communication links between nodes have randomness and instability, which directly affect the network simulation evaluation. In the previous research of wireless sensor networks, Focusing on protocol and energy consumption, the channel design is mostly based on the ideal channel model. COOJAN Contiki OS Java) as a newly developed wireless sensor network simulation platform in recent years, it can efficiently simulate hundreds of nodes of wireless sensor networks. However, the channel model can not well reflect the fading characteristics of wireless channel in real environment. It is very important to develop the channel model suitable for wireless sensor network application environment. Firstly, this paper introduces the research progress and application of wireless channel model, and analyzes the shortcomings of channel model in COOJA. The possibility of developing and integrating new channel models is put forward, and the classical channel models based on the statistical law of measured data are studied. In this paper, the channel models suitable for wireless sensor network application environment are studied and designed from the aspects of large-scale fading and small-scale fading respectively. When the channel exists the line-of-sight propagation, the received signal is distributed according to the Rice. Combining with the Clarke model and the simplified idea of Jakes, the Rice channel model is constructed. The simulation results show that the proposed method has the statistical characteristics of the slave Rice distribution. The large scale fading is based on logarithmic path loss, the Plane-Earth model is adopted and compared with the measured double slope model in different environments. Finally, the channel for five different environments is designed: flat cement land, flat sand land, flat grassland, low bush, forest environment, etc. However, the Plane-Earth model can not be described uniformly in the latter two environments because of the influence of the presence or not of the line-of-sight component, so the latter two environments use a single slope model. Finally, based on the in-depth study of the architecture and simulation mechanism of the large scale simulation platform COOJA, Five channel models are designed and implemented by using Java language. The functions of signal strength prediction and packet error rate calculation are completed. The channel model is integrated into the COOJA simulation platform and compared with the original channel. In comparison with the original channel, the prediction value of the original channel prediction model is quite different. In terms of simulation performance, through the establishment of a 500-node wireless sensor network simulation scene, due to the higher computational complexity of the new channel model, the overall simulation efficiency is slightly reduced, but the channel model prediction is more reasonable. One of the channel models that can be used in future COOJA wireless sensor network simulation.
【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】