Uma solução de inferência de culpabilidade em cenários de acidentes de trânsito por meio de vanets

Abstract

Traffic accidents are common facts amid the large number of vehicles fighting for space on public roads. It can be observed that questions directed to who is to blame for the collision also became common generating costs due to the need for technical expertise at the accident sites, legal disputes and various other disorders. Taking this reality into account, this paper seeks to perform an analysis and create a solution that assists in solving these problems, pointing out the cause of the collision. Once the guilt inference is made, the means of communication are presented in VANETs to send the data to ensure that the information is sent, regardless of where the vehicles may be. Based on forensic methods that allow accurate reconstruction of rear and side collisions, algorithms are defined to infer guilt in these scenarios using data generated by simulators. Observing the same scenarios is made an analysis of expert opinion, which perform an evaluation through images, videos and multiple choice questionnaire. A comparison of the results obtained is performed. Through comparison it can be seen that the inference of guilt with accurate information is possible through the proposed solution. It can also be verified, by applying the algorithms, that there are more important factors than others in defining blame in different types of collisions.