Due to the need for road safety, Vehicular Ad-hoc NETworks (VANETs) field is developing rapidly to improve the reliability and safety of road traortation. Despite numerous innovations in manufacturing technology and improvements in vehicle design, avoiding road accidents and safety of passengers is still the main concern in road traortation. However, presence of misbehaving agents in the network who send false information to achieve malicious goals, threatens the efficiency of the network. In these circumstances, using encryption mechanisms alone is not enough, because the misbehavior may be done by vehicles which are properly authenticated. Therefore, it is important to present a proper model for trust establishment in such networks. In this thesis, we review some well-known models in trust. Moreover, common attacks in trust context are discussed. Then a distributed model is presented to evaluate the trust. In this model, trustworthiness of the received messages are estimated with the help of trust assessment for vehicles. It defines the dynamic threshold to collect the required messages about the events to take decisions in regard to the maximum available evidences. The proposed model is based on roles and events and considers sparsity which is a problem in VANETs. Simulation results show that the proposed model presents high detection accuracy in the face of real and bogus events and its accuracy is acceptable in both sparse and dense networks. Keywords: Vehicular Ad Hoc Network, Trust evaluation, Security, Safety messages
