Vehicular Social Network (VSN) emerges as one meaningful scenario of Mobile Social Network (MSN) and aims at addressing the traffic congestion and security problems in smart cities. However, the current IP based network connection cannot efficiently solve the high mobility and dynamically changing topology issues caused by VSN, which results in long routing latency and low cache hit ratio. In this regard, this work intends to optimize the routing latency and the cache hit ratio by solving the content delivery problem in VSN, during which the concepts of fog/edge computing and Information-Centric Networking (ICN) are introduced, since they can provide a new perspective of network connection in VSN by concentrating on contents instead of IP addresses. Firstly, the ICN routing table structure is expanded for the purpose of providing key routing basis for both interest and data packets. Secondly, the vehicle social information is fully explored from the points of relationship strength and interest degree to offer reliable data foundation for community detection and packet forwarding. Thirdly, the dynamic vehicle communities are constructed based on an improved community overlap propagation algorithm. By monitoring the content popularity and cache distribution in the dynamic communities, this paper proposes the vehicle-adaptive caching scheme which guarantees the cache diversity and improves the packet delivery rate. After that, a vehicle-adaptive interest packet routing scheme is proposed based on relationship strength and interest degree for both inter/intra communities. The experiment results indicate that the proposed mechanism can improve the network performance in terms of the packet delivery rate, the average hop, the average delay and the overhead.

车载社交网络（VSN）成为移动社交网络（MSN）的一种有意义的方案，其目的是解决智能城市中的交通拥堵和安全问题。但是，当前基于IP的网络连接无法有效解决由VSN引起的高移动性和动态变化的拓扑问题，从而导致较长的路由延迟和较低的缓存命中率。在这方面，这项工作旨在通过解决VSN中的内容传递问题来优化路由延迟和缓存命中率，在此期间引入了雾/边缘计算和信息中心网络（ICN）的概念，因为它们可以提供通过专注于内容而不是IP地址，VSN中的网络连接有了新的视角。首先，为了提供兴趣和数据包的关键路由基础，扩展了ICN路由表结构。其次，从关系强度和兴趣程度两方面全面探索车辆社交信息，为社区检测和报文转发提供可靠的数据基础。第三，基于改进的社区重叠传播算法构建动态车辆社区。通过监控动态社区中内容的流行度和缓存的分布，提出了一种车辆自适应缓存方案，该方案可以保证缓存的多样性并提高数据包的传输速率。在此基础上，提出了基于车内/车内社区的联系强度和兴趣度的车辆自适应兴趣包路由方案。