SEGMETRIK: UM PROTOCOLO PARA SEGMENTAÇÃO E RASTREAMENTO DE DESEMPENHO NA DESCOBERTA OPORTUNÍSTICA DE PONTOS DE INTERESSE EM VANETS

• Ano de defesa: 2018
• Autor(a) principal: Daniel Dener Assis de Sousa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais
• Palavras-chave em Português: Redes de computadores Protocolos; Computação; Redes de computadores
• Link de acesso: https://hdl.handle.net/1843/ESBF-B94EVT

In Intelligent Transport Systems (ITS), Vehicular Ad Hoc Networks (VANET), a subclass of Mobile Ad Hoc Network (MANET), play a key role in the large-scale deployment and deployment of security and non-security applications. Among applications that promote entertainment, an area that has recently attracted attention and been investigated is the Point of Interest (POI) discovery and location service. Such POIs as restaurants, hotels and others could benefit from these networks to disseminate their ads in real time to a large number of vehicles. In the same way, potential consumers can use these applications to find products and services during their routes within cities as well as on highways. Commercial advertising campaigns, especially sponsored ones, need to have features that increase their effectiveness, such as metrics and performance statistics. Ad tracking is the general practice of collecting user profile data and your interaction with ads to optimize their effectiveness. A common metric of ad performance is CPC (Cost Per Click), which indicates the average price paid by the advertiser for each click resulting from user interaction with the ad. CPC is influenced by ad efficiency and impression price. An ad impression is the time when a user sees the ad. To improve ad efficiency, even if the user sends a specific request, advertisers often try to match ads to potential customers so they do not waste impressions - and financial resources - on customers for whom that advertisement is not relevant. This matching, known as targeted ads in online marketing, takes as input some representation of the user's profile in terms of measurable observations, such as location, momentary situation (date and time), and consumption history. In this work we present SEGMETRIK, a low overhead application protocol that satisfies the requirement of ad targeting and tracking. There are already proposals in the literature for the use of applications aimed at the dissemination of location-aware information and delivery of advertisements in vehicular networks. However, these papers do not present a lightweight, low overhead application protocol, with ad targeting and tracking of ad performance on opportunistic POI discovery services and compliant with VANET standards. The solution proposed in this work includes a series of features and characteristics of which do not exist concomitantly in published works of our knowledge, which makes this a unique approach. This dissertation presented simulations and analysis in two scenarios: (i) partial ad targeting using V2V communications, involving a POI and 505 vehicles in a 14 km2 area based in the city of Manhattan and (ii) complete ad targeting using V2I communications, involving 5 RSUs, 15 POIs and 192 vehicles in an area of 10 km2 based in the city of Belo Horizonte. In this way, this study demonstrated through simulations and analysis that ad targeting and performance tracking strategies can be applied successfully to applications that use opportunistic discovery of points of interest in vehicular networks.