This thesis has been dedicated to addressing the misbehaviour detection problem in vehicular networks. Specifically, we focus on two major issues in PHY layer and application layer respectively: Radio Frequency (RF) Jamming attacks and Sybil attacks. Specifically, we adopted three different machine learning methods including Distance based clustering, Support Vector Machine (SVM) and k-nearest neighbours (kNN) in Sybil nodes detection. Based on variation between benign vehicles and Sybil nodes in their driving patterns, the non-existent virtual nodes can be detected. For RF jamming attacks, we focused on the design of countermeasure for the control channel jamming issue in vehicular networks, which is of vital importance to the safety of I2V communications. We proposed to adopt the cooperative relaying techniques to address the control channel jamming problem in vehicular networks, which is based on the idea that the vehicles outside of the jamming area can serve as relays to help forward the control channel signal to the victim vehicles through other the jamming-free service channels. Thus, we extended the jamming issues in multi-antenna RSU scenarios, where the RSU can serve multiple groups of vehicles simultaneously using the multi-group multicast beamforming technique. As a solution, we propose a two stage anti-jamming scheme, whereby the vehicles who have successfully decoded the signal received in the first stage will be selected as relays to cooperatively serve the victim vehicles in the second stage using the coordinated beamforming techniques over a jamming-free service channel.