【作者】 高峰；

【导师】 夏克文；

【作者基本信息】 河北工业大学 ， 微电子学与固体电子学， 2011， 硕士

【摘要】 无线传感器网络具有节点尺寸小,能够感知、处理数据,低功耗、低花费等特点,利于在复杂环境下采集、处理大量数据,但其存在硬件资源和电池容量有限,网络拓扑结构复杂,不易更换电池的缺陷,为解决这一问题,目前提出了许多高能效的信息融合算法和网络通信协议,其中,自然计算是以自然界机理为设计基础的算法,适用于解决那些难以建立有效的形式化模型而无法处理的问题,在大规模复杂系统的优化设计、优化控制等领域具有广阔的应用前景。本论文主要对无线传感器的网络模型进行建模,并应用自然计算的方法进行仿真优化设计,具体研究内容如下:(1) WSN路由模型研究。首先介绍对WSN的基本体系结构和网络覆盖形式,然后阐述了WSN中具有代表性的路由协议,并研究了链状和簇状路由模型。对于这两种路由模型,从能量的角度进行了分析和建模,建立了能量优化公式,提出解决方案。(2)自然计算(NC)技术研究。主要对群智能算法和云智能进行研究,对于蚁群算法和粒子群算法存在的不足进行相应改进。对于链状路由模型,提出了基于自适应蚁群算法(SA-ACO)算法的无线传感器网络能量的优化方法,它包括动态概率选择、优化信息素矩阵和遗传变异等过程;对于簇状路由模型,提出了基于云自适应粒子群算法(CAPSO)的无线传感器网络能量的优化方法,它主要包括网络分簇、网络能量模型建立、云PSO算法迭代优化等过程。(3)基于自然计算的WSN仿真设计研究。基于SA-ACO算法的路由算法,以自适应的方法优化信息素参数,通过对链状路由模型的仿真,表明其降低了WSN节点的网络能耗,提高了网络节点的生存能力,并且引入一定得变异规则,从而提高了算法的搜索能力,防止了节点陷入局部最优解,通过与传统ACO算法比较,该算法网络能耗和节点死亡率较低。基于CAPSO算法的路由算法,结合了云智能模糊性和随机性相结合的特点,改变了基本PSO算法惯性权重固定取值的方法,将粒子根据不同的适应度情况采用不同的惯性权重,并应用云模型进行优化,结果表明CAPSO算法有效降低了网络能耗和网络延时。更多还原

【Abstract】 Wireless sensor networks(WSN) can sense, process data and has low power, low cost, nodes of small size,which conducive to collecting and processing large amounts of data in the complex environment. But it has some shortcomings, such as limited hardware resources and battery capacity, complex network topology and difficult to replace the battery. To solve this problem, many of energy-efficient data fusion algorithm and network communication protocols have been proposed. Natural Computation(NC) is based on natural mechanism for solving those difficult to establish effective formal model and deal with it. In large-scale optimization of complex systems design, optimal control and other fields, it has broad application prospects.This thesis focuses on WSN modeling, and applying the method of natural computation for simulation and optimization design.The specific contents are as follows:(1)Research on WSN routing model. Firstly, the basic architecture of the WSN and the form of the network coverage are introduced.Secondly, representative WSN routing protocols is elaborated, and the chain and cluster routing model are studied. For the two routing model, from the point of view of energy analysis and modeling, energy optimization formula is established and solutions is proposed.(2) Research on Natural Computing(NC).The swarm intelligence algorithm, and cloud intelligent algorithms are mainly studied and shortcomings of the ant colony algorithm and particle swarm algorithm are improved. For the chain Routing model, an optimal approach for WSN is proposed based on self-adaptive ant colony optimization (SA-ACO) energy optimization,which includes the dynamic probability selection, optimizing pheromone matrix and genetic variation of combining process;for the cluster Routing model, an optimal approach for WSN is proposed, based on the Cloud Adaptive Particle Swarm Optimization (CAPSO) algorithm, which includes network clustering, network modeling, iteration optimization with CAPSO algorithm, and so on.(3) Research on WSN Model with Natural Computation. The routing algorithm based on SA-ACO optimizes pheromone parameters through adaptive approach. The simulation for chain routing model prove that WSN node energy consumption is reduced, the viability of network nodes is increased. And the introduction of a certain variation of the rules is to improve the search ability of the algorithm to prevent the node into a local optimal solution. Compared with the traditional ACO algorithms, the algorithm has a lower energy consumption and node mortality. The routing algorithm based on CAPSO combined with fuzziness and randomness of cloud intelligent algorithms, change the basic PSO algorithm which has a fixed value of inertia weight. According to the fitness of different particles, the algorithm uses different inertia weight, and is optimized by the cloud model. Results show that, CAPSO reduces energy consumption and network delay.更多还原