【作者】 马军营；

【导师】 李廉；

【作者基本信息】 兰州大学 ， 软件工程， 2016， 硕士

【摘要】 因为泥石流灾害具有流量大、流速快、毁坏力强等特点,导致一直以来对泥石流的监测都缺乏切合实际的方案。伴随着无线传感器技术的出现和发展,以其低成本、低功耗、快速响应等优点,为我们在泥石流临灾预警和监测方面提供了新的解决思路,泥石流临灾监测的无线传感器网络应运而生。然而在泥石流灾害爆发时,复杂的电磁环境和多变的信道条件导致了我们必须面临的关键问题——如何确保无线传感器网络数据的可靠传输、实现网络节点的能量负载均衡以期最大限度地延长网络的寿命。为了解决上述的关键问题,我们将协作通信技术应用于泥石流临灾监测的无线传感器网络中,并对三节点协作模型、含S-D链路协作模型和无S-D链路协作模型进行了深入探讨,同时对放大转发协议(AF)、解码转发协议(DF)、选择转发协议(SDF)和增量中继协议(IF)进行了研究,并在IF协议的基础上对多中继无线传感器网络模型、单中继无线传感器网络模型以及单信道直接传输信道模型进行了节点功耗和中断概率的对比分析,证明了多中继协作传输在泥石流临灾监测环境下具有独特的优势。但是上述协议都存在有共同的缺陷——相对于整个无线传感器监测网络来说,把中继链路或者传感器节点能耗以单独的部分进行分析和研究,而没有从整体上对网络的功耗以及节点的能量负载均衡进行把握。为此在基于IF协议的多中继协作模型下,我们提出一种依托多种网络性能指标并且基于联合预测的能量负载均衡多中继选择策略,并应用于泥石流临灾监测的无线传感器网络模型中,实现了提高网络的通信质量、节点能量负载均衡的目标。基于联合预测的能量负载均衡多中继选择策略的关键技术有:一、基于S-R信道的LQI和中继节点的剩余电量形成协作传输小组的雏形。二、利用马尔可夫预测方法计算出雏形协作传输小组的中断概率Pout,如果Pout<Pout-threshold那么就会把雏形协作传输小组作为最终确定传输信息的协作小组;否则,丢弃当前雏形协作传输小组并进行重新生成。三、整个基于联合预测的多中继选择策略以“轮询”的方式执行,直至整个多中继协作无线传感器监测网络生命的结束。通过实验得出,我们提出的基于联合预测的多中继选择策略使得监测网络的中断概率、误码率都有明显降低,并且能量的消耗能均匀的分布于所有的中继节点。更多还原

【Abstract】 Due to the debris flow has the characters with large mount,high flow speed and strong damage, it has been the result of a lack of monitoring of appropriate solution. With the emergence and development of wireless sensor technology, it provides us with new ideas to solve the mudslides temporary disaster monitoring and early warning, which has the advantage with low cost, low power consumption and fast warning. However, the complex electromagnetic environment and changing channel conditions always company with the occurring of debris flow, there are new key problems in a temporary disaster debris flow monitoring system for wireless sensor networks,which are how to guarantee the reliable transmission of sensor data and achieve the load balancing of sensor nodes.In order to solve the problems had been mentioned in the wireless sensor network of temporary disaster debris flow monitoring system, we have made in-depth discussion on the classic collaboration model, including S-D collaboration model and none S-D collaboration model, while AF protocol, DF protocol, SDF protocol and IF protocol were studied. Finally, we made a comparison with multiple relay collaboration model,a single relay collaboration mode and single-channel direct transmission based on IF protocol. We know the multiple relay collaboration protocol has the best performance with the outage portability and energy consumption in the face of disaster debris flow monitoring environment. However, the existing agreement have a common flaw in regarding the relay link or the node’s energy consumption as a separate part when making the analysis and research. We must pay more attention to the whole network on outage probability and the load balancing of sensor nodes.To do this in a multi-relay cooperation model based on IF protocol, we propose s multi-realy cooperation joint strategy based on prediction, and apply the strategy to the mudslides temporary disaster monitoring multi-relay cooperative wireless sensor networks model. We achieve the goal of improving the quality of communication network and guaranteeing the load balancing of the whole network. The main idea of the predictive multi-relay cooperative joint strategy including three parts. First, based on the remaining energy of sensor node and the LQI of the S-R link, the prototype of cooperative transmission channel group is formed. Second, using the Markov forecast method to calculate the prototype group’s cooperative transmission outage probability Pout, if Pout<Pout-threshold then it will put the prototype of a cooperative transmission group as a collaborative group to finalize the transmission of information; otherwise, discard the current prototype coordinated transmission group and reselect the group. Finally, the multi-relay cooperative joint strategy based on predicted will be used in the manner of "poll", until the end of the entire multi-relay mudslides temporary disaster monitoring system in collaboration. The simulation results which have been obtained,shows that the multi-relay cooperative joint strategy based on prediction, have significantly reduced the error rate and outage probability of the network. Also in the simulation process we find that when a relay node is died in the network, the other following nodes are also close to the energy depletion edges.更多还原