Cooperative Socio-aware Dynamic Backoff Optimization for Urban VANETs

The communication graph formed by vehicles is an abstract and dynamic graph that faces intermittent disconnections due to the high mobility of vehicles. This graph is governed by traffic signaling, urban planning and drivers behaviours. Even if Social Network Analysis (SNA) proved its benefits in the study of abstract social networks, it poses new challenges when applied to dynamic networks such as vehicular ad hoc networks (VANETs). This paper aims at studying how social centrality and community metrics, two important social properties in SNA can help to mitigate the poor performances in vehicular networks. This paper first explores the possible correlation that exists between performance degradation and social metrics in urban-like maps. After conducting an extensive quantitative analysis on the correlation between social and underlying networking properties of vehicular networks under different densities, space clustering coefficient, transmission range, we observed that closeness centrality, clustering coefficient and lobby index can help to identify special vehicles - those vehicles that have a higher opportunity to access the channel than others - and discuss the implications of such findings in the context of a dynamic backoff optimization for the IEEE 802.11p MAC layer.