移动物联网是由射频识别、无线数据通信技术、无线互联网等技术所构成的网络体系，同时又是移动通信、移动终端、物联网三大领域碰撞和融合的产物，并成为了物联网发展的重要模式和途径。近年来，智能入网设备数量持续激增，尤其是移动智能终端的迅速普及，为恶意节点提供了更多网络攻击的机会，再加上自私用户的不合作行为、终端的频繁移动或有限能量易丧失、网络规模的不可预知等特性，为当前移动物联网服务的连续性、可靠性、节能效果等方面带来了一系列挑战。因此，本文基于移动物联网的网络体系架构，围绕受移动终端影响较大的移动物联网业务中的数据转发及路由问题，针对移动接入网的可信性转发模型、融合多层网络的可信性转发模型以及基于移动Ad Hoc网络(MANET)的低能耗路由机制进行了深入的研究，主要的研究工作包括以下三个方面： (1)针对网络中因恶意或自私节点破坏或不转发传输数据所导致的隐私泄露或服务中断的问题，提出了基于节点多元社交属性的移动接入网可信转发模型。该模型首先根据实际移动接入网应用中对象间不对等的信任关系，利用联系概率、服务度、新近性对移动接入网中节点间局部的单向可信性关系进行了定义；然后，再结合节点的度，利用向量内积和调整因子分别对节点间社交相似性和节点的个体中心性进行度量，从而获得节点间全局的双向可信性关系；最后使用万有引力定律集成双向可信性因子，实现对节点间转发能力的评估。基于两个真实数据集的实验证明，所提出的模型具有较高的稳定性，且在兼顾数据累积投递率的同时，降低了服务的平均投递时延和平均网络代价。 (2)针对包含多种网络节点类型且完整网络规模不可预知的移动物联网数据跨空间可靠传输问题，提出了基于雾计算模式的混合网络可信转发模型。该模型首先针对由移动终端和雾基础设施组成的混合网络建立节点间的单向可信性关系，并利用种子扩展和随机游走优化算法构建一个可信重叠社团动态检测模型，对具有不完整性结构信息的混合网络进行基于社团的动态分割；其次，根据雾基础设施节点的远距离通信功能，依照一定的规则对所有连通的雾基础设施节点进行分割，形成多个不连通的子图；然后，基于雾基础设施节点的这些子连通分布，对第一步中得到的社团根据设定规则进行重新联合，形成跨空间的逻辑联合边缘社团结构；最后，在逻辑联合边缘社团结构的基础上，分别对移动节点和雾基础设施节点建立对应的转发能力度量，并设置相应的转发规则。基于两个真实数据集的实验证明，所提模型在数据累积投递率、平均投递时延和平均网络代价三个方面均超越了一些流行的转发模式。 (3)针对移动物联网环境下因MANET中移动终端间通信连接频繁丧失引起的服务再搜索再路由所导致的能耗大量增加的问题，提出了一种动态Cloudlet辅助的MANET节能路由机制。该机制首先为每个移动终端建立相应的临时文件用于记录其身份识别码和一定时间段内的路由信息，并考虑使用D2D(Device-to-Device)作为移动终端间的主要通信方式；然后，利用Cloudlet作为小型数据中心，为所有Cloudlet创建一个服务共享关系表并设置Cloudlet间的合作机制；最后，不管移动终端如何变换位置，依靠上述设置，它们总能通过其他途径快速地再搜索再路由到已丧失连接的服务。通过使用和不使用所提机制两种情况的多次实验对比结果证明，使用所提机制大大降低了MANET中移动终端再搜索再路由服务的能量消耗。 利用节点的多元社交属性为移动接入网构建的可信转发模型，通过强化节点间交互关系的可信性，能够尽量避免恶意或自私节点为转发带来的不利影响，具有良好的可扩展性；基于雾计算模式的混合网络可信转发模型，在网络规模不可预知的情况下，结合动态社团结构和网络边缘固定设施的跨空间长距离转发能力，有效地完成了移动物联网数据在局部范围和远距离的可靠传输；利用Cloudlet间的动态协作机制能大大缩短MANET中移动终端再路由或再搜索服务的时间，在提高服务运行效率的同时，还降低了系统能耗。总之，以上三个机制的应用可以实现数据高效率、低能耗、安全可靠的传输，能够为追求可持续性发展的移动物联网应用提供最基本的保障。 关键词：移动物联网；数据转发；可信性；路由；节能

 Mobile Internet of Things (IoT) is a network system composed of radio frequency identification, wireless data communication, wireless Internet technology and other technologies. It is also the result of collision and integration of mobile communications, mobile terminals and IoT, and it is an important mode and way for the development of IoT. In recent years, the number of intelligent access devices continues to increase sharply, especially the rapid popularization of mobile intelligent terminals, which provides more opportunities for malicious nodes to attack the network. In addition, the non-cooperative behavior of selfish users, the frequent movement of terminals or the loss of limited energy on terminals, unpredictable network size and other characteristics, bring enormous challenges for the continuity, reliability and energy-saving effects of current mobile IoT services. Therefore, based on the network architecture of mobile IoT, focusing on the data forwarding and routing problems in mobile IoT bussiness which are greatly affected by mobile terminals, this paper makes a deep research on the trustworthiness forwarding model of mobile access network, the trustworthiness forwarding model fusing multi-layer networks and the energy-saving routing mechanism based on mobile Ad Hoc network (MANET). The work includes the following three aspects: (1)To solve the problem of privacy leak or service interruption caused by malicious or selfish nodes destroying or not forwarding data, a trustworthiness forwarding model of mobile access network based on multivariate social relationships between nodes is proposed. First, according to the unequal trust relationships between objects in the actual scenarios, this model first defines the local unidirectional trustworthiness relationships between nodes by using the contact probability, service degree and recency between objects. Then, combining the degree of nodes, the social similarity between nodes and the individual centrality of nodes are measured respectively by using the product of vectors and the adjustment factor, so as to the global two-way trustworthiness relationships between nodes. Finally, the law of universal gravitation is used to integrate the two-way trustworthiness factors to evaluate the forwarding capability of nodes. Experiments on two real data sets show that the proposed model has high stability and reduces the average delivery delay and average network cost while taking into account the cumulative delivery rate of data. (2)To deal with the problem of reliable data transmission across space in large-scale mobile IoT when there are many types of network nodes and the complete network structure information can not be known beforehand, a trustworthiness forwarding model of hybrid network based on fog computing is proposed. First, the model establishes the unidirectional trustworthiness relationships between nodes in a hybrid network composed of mobile terminal nodes and fog infrastructure nodes, and constructs a dynamic trustworthiness overlapping community detection model by using seed expansion and random walk optimization algorithm, which can dynamically segment large-scale mobile IoT based on communities; Second, considering the long-distance communication function of fog infrastructure nodes, all connected fog infrastructure nodes are partitioned according to certain rules and form multiple sub-connected components. Third, based on the distribution of sub-connected components, the communities obtained in the first step are re-united according to the set rules to form logical joint edge community structures. Finally, on the basis of the logical joint edge community structures, the corresponding forwarding capability metrics and the corresponding forwarding rules are established for mobile nodes and fog infrastructure nodes, respectively. Experiments on two real data sets show that the proposed model surpasses some popular forwarding modes in terms of cumulative data delivery rate, average delivery delay and average network cost. (3)To solve the problem of increasing energy consumption caused by re-searching and re-routing services while link breakages between mobile terminals in MANET, a dynamic cloudlet-assisted energy-saving routing mechanism is proposed. First, the mechanism establishes temporary files for each mobile device to record its ID and routing information in a certain period, and considers using D2D (Device-to-Device) as the main communication mode between mobile terminals. Second, it uses a cloudlet as a small data center, creates a service sharing table for all cloudlets and establishes a cooperative mechanism between cloudlets. Finally, no matter how mobile terminals change their locations, relying on the above-mentioned settings, they can always re-search and re-route the lost services quickly through other routes. The experimental results show that the proposed scheme greatly reduces the energy consumption of re-searching and re-routing services for mobile terminals in MANET compared with not using the proposed mechanism. The trustworthiness forwarding model constructed for mobile access network based on multi-social attributes of nodes can avoid the adverse effects of malicious or selfish nodes on forwarding by enhancing the credibility of interaction between nodes, and has good scalability. In the case of unpredictable network size, the trustworthiness forwarding model of hybrid network based on fog computing combines dynamic community structure with trans-space long-distance forwarding ability of fixed facilities on network edge, and effectively completes the reliable transmission of mobile IoT data in local scope and long-distance. Using the dynamic cooperation mechanism between Cloudlets can greatly shorten the time of re-routing or re-search service for mobile terminals in MANET, and improve the efficiency of service operation, but also save a lot of system energy. In a word, the application of the above three mechanisms can achieve high efficiency, low energy consumption, safe and reliable transmission of data, and provide the most basic guarantee for the sustainable development of mobile IoT applications. KEY WORDS: Mobile Internet of Things; Data Forwarding; Trustworthiness; Routing; Energy Saving