基于动态社会感知的移动机会网络群组通信研究

Mobile opportunistic networks, that forward messages through the contact of mobile nodes, has been widely adopted in mobile social networks, VANETs, mobile ad hoc networks etc., which have attracted considerable attentions in the research and industry community. In such networks, opportunistic mobile group communication has been pervasive in a wide range of scenarios such as disaster rescue, data push, community information dissemination etc., which makes multicast routing a critical building block of the design of the network. Nevertheless, these networks suffer from limited node buffer and energy supply, as well as selfish behaviors. To address these problems, this proposal first plans to build models to capture the essential dynamic social characteristics that evolve along time and space. Based on that, to address the limited buffer and incomplete transmission problem due to transient contacts, cooperative buffer management strategies between nodes forwarding fragments of same messages will be designed. Next, power consumption will be modeled considering the dynamic social characteristics and use optimization techniques to derive the optimal or suboptimal solutions. Based on that, the power balancing algorithm will be studied and proposed. The selfish issues will also be studied, and rewarding and punishing mechanisms will be proposed accordingly. We introduce the game theory to model the game between group nodes as "cooperation within a group, and competition among groups". Equilibrium points under different incentive mechanisms will be derived to verify the efficacy of these mechanisms. With the insights from the game model, a joint reputation mechanism will be proposed to promote the cooperation of the multicast nodes. At last, a comprehensive multicast protocol and system prototype will be developed integrating all the above components, as to boost the comprehensive performance of dynamic social-aware multicast in mobile opportunistic networks.

基于节点机会式接触通信的移动机会网络被广泛地应用于移动社交网络、车载智能网络、移动自组网等场景，具有重要的研究意义。移动机会网络群组通信在抢险救灾、数据推送、社区消息分发等领域具有广阔的应用价值，其相应的组播路由设计是保证群组通信的关键问题。但是移动机会网络节点缓存和能量有限，自私行为等问题给组播路由设计带来了挑战。本课题首先分析随时空变化的动态社会特性感知模型。基于该模型，针对数据无法在短暂连接时间内完成传输及节点缓存有限问题，设计分片数据转发节点间合作缓存组播路由策略；同时利用动态社会感知模型，建立节点能量消耗最优化模型，推导最优或次优解，设计能量均衡组播路由策略；最后，对自私性问题设计抑制/激励机制：引入博弈论对节点 “组内合作，组间竞争”行为进行建模，推导不同激励机制下的均衡点，建立联合声誉机制促进节点合作。最终实现移动机会网络基于动态社会感知的组播路由通信协议和系统原型。

移动机会网络被广泛地应用于移动社交网络、车联网、移动自组网等场景，具有重要的研究意义。移动机会网络群组通信在抢险救灾、数据推送、社区消息分发等领域具有广阔的应用价值，其相应的组播路由设计是保证群组通信的关键问题。本项目研究移动机会网络节点缓存和能量有限，自私行为等问题给组播路由设计带来了的问题。首先分析当前研究现状，包括缓存管理、群组安全策略、绿色能量优化、动态感知方面的研究前景，提出本课题的技术思路；分析随时空变化的动态社会特性感知模型，结合能量消耗模型提出能量优化和降低负载开销的组播路由策略；基于动态社会感知模型，设计社会特性考虑的组播路由缓存管理策略；对自私性问题设计抑制/激励机制，通过衡量节点信任度，设计抑制自私性行为组播安全策略；并基于Android平台，实现了自私性安全策略和能量消耗考虑的移动机会网络组播路由通信系统；进一步研究LEO低轨卫星通信网络中路由负载均衡问题，基于卫星网络移动特性和缓存信息，提出了动态分配通信流量的分布式负载均衡路由策略；利用实际跟踪数据，分析车联网的移动特性，构建多目标数学模型，提出优化的车联网边缘计算服务器放置方案。总之，本项目对移动机会网络组播路由相关的缓存管理、群组安全策略、绿色能量优化等领域进行了研究，实现了安全和能量考虑的移动机会网络组播通信系统，为移动机会网络的性能优化提供了理论基础和实际应用平台，有力推进移动机会网络组播通信技术的扩展和推广。