Dissemination Of Emergency Warning Messages In Vehicular Social Networks

With the immense increase in the number of vehicles,road accidents have become one of the leading causes of fatalities around the world with about 1.35 million deaths and a loss of 3%global GDP per year.However,several studies indicate that this risk can be significantly reduced with a prior collision avoidance system.Therefore,vehicular communication-based road safety solutions such as the issuance of Emergency Warning Messages(EWMs)have been a major research attraction recently.However,various problems such as broadcast storm and hidden terminal,connectivity,reliability,and authenticity issues still persist in conventional vehicular networks reducing their efficacy.Therefore,this thesis explores new avenues such as Vehicular Social Networks(VSNs)that offer higher reachability,flexibility,and scalability to mitigate the aforementioned issues with three main contributions that are described as follows:First,this thesis proposes a novel Social Utility-based Dissemination Scheme(SUDS)for EWMs in VSNs exploiting social network-based properties such as interests,centrality,and friendships to mitigate broadcast storm and hidden node problems.For this purpose,SUDS employs hybrid architecture of VSNs solving connectivity and contact-duration related issues in sparse and high mobility environments along with efficient algorithms for centralized and distributed dissemination.Extensive experiments with a varying range of vehicular density,speed,and distance for network measuring parameters such as delivery ratio,delay,and overhead indicate that SUDS outperforms state-of-the-art dissemination schemes with high accuracy.Second,to solve the issue of false alarms such as fake EWMs triggered by malicious users that can precariously disrupt vehicular traffic especially on highways,this thesis proposes a novel Trust based Dissemination Scheme(TDS)for EWMs in VSNs.TDS ensures the authen-ticity of EWMs by exploiting the user-posts interaction and credibility networks to identify the true nature of generated EWM.For this purpose,TDS incorporates a comprehensive reputation mechanism by calculating a trust-score for each node based on its historical behavior,social-utility,and contribution.The performance of TDS remains significantly accurate even in the presence of a very high density of malicious nodes at a varying range of vehicular speed and density.Moreover,extensive experiments under different network conditions and parameters validate the reliability,robustness,and scalability of TDS in VSNs.Third,this thesis addresses the issues of selfishness,connectivity,limited network re-sources,and their adverse effects on real-time dissemination of EWMs in VSNs by proposing a novel Cooperative Real-time Dissemination(CRD)scheme using crowdsensing and social intelligence to differentiate between a selfish and a cooperative node.For this purpose,CRD calculates a new tie-strength metric based on intimacy,similarity,and interactions.Moreover,CRD incorporates a recursive evolutionary algorithm for each node's reputation calculation and update,which is used to estimate its state-transition probability for selection as a super-spreader.Finally,CRD's performance is evaluated in terms of network parameters such as delivery-ratio,delay,hop-count,and message-overhead for varying range of vehicular density,vehicular speed,and density of selfish nodes based on realistic vehicular mobility traces.In addition,comparative analysis of CRD's performance indicates significant superiority over state-of-the-art dissemina-tion schemes in VSNs for the aforementioned network parameters.