A Targeted Advertising System For Vehicular Networks

The emergence of wireless communication and self-driving automobiles has drawn great attention to Vehicular Ad Hoc Networks(VANETs),where vehicle nodes can interact with each other through wireless communications.A variety of applications built on VANETs have been developed such as monitoring real-time traffic,sharing information like music and movies,and advertising over VANETs.Advertising over VANETs is mainly based on Social Internet of Vehicles(SIoV).Communications between vehicles and between vehicles and roadside infrastructures form SIoV,and advertisements dissemination around the city is conducted through the formed SIoVMerchants want to achieve a largest benefit through ads with a limited budget.The maximization of advertising benefits is mainly determined by the influence of the advertisement and potential benefits generated by users receiving advertisements.The former factor mainly corresponds to an advertisement dissemination process,while the latter factor corresponds to an advertisement recommendation process.At the same time,the targeted and accurate recommendation process of advertisements can also improve users’ experience,increase users’ acceptance of advertisements.This paper focuses on the above two processes of the advertising dissemination system,and researches the targeted advertising distribution system build on VANETs which combines with the social networks and collaborative filtering recommending systems.Taking characteristics of VANETs and advertisement information into account,this paper maximizes the spread of advertisements and potential benefits of advertisements after recommendation.During the dissemination process,this paper first proposes a model for advertising information dissemination of VANETs,which is based on the information dissemination model in social networks.In this proposed model,advertisements are forwarded by selected "seed vehicles".Mathematical abstraction is carried out to abstract the problem of maximizing the influence of advertising.It is proved that the proposed problem is NP-hard To approximate the problem,this paper proposes a mobility-based "seed vehicles"selection algorithm.Experimental results implicit the proposed algorithm can reduce computation as well as achieve a large coverage of ads.At the same time,a targeted recommendation system model for advertising is proposed.Along with a proposed buffering scheme,a recommending algorithm for advertising in VANETs is proposed based on collaborative filtering recommending systems.According to the advertisement characteristics and the mobility of vehicles,the algorithm uses "tags" to recommend while adding a vehicle location information when determining neighbor nodes.Experimental results prove that the proposed algorithms can achieve an excellent performance.The innovations are proved to be effective as well.This thesis consists 5 chapters,including 14 figures,7 tables and 52 references.