Novel vehicle crash risk detection based on vehicular sensory system

Vehicular on-board devices have been widely applied for enhancement of perceiving driving status and road traffic environment, which in turn, can be used to detect vehicle crash risk and provide drivers with pre warning or safe driving assistance in dangerous situation. This paper aims to propose a vehicular cyber physical system and its logical framework for crash risk identification and warning system based on vehicular on-board devices. Wireless communication devices connecting subject vehicle to approaching vehicles (V2V) and vehicles to roadside equipment (V2I), are applied to comprehensively track the motion status of the moving vehicles and road target (approaching vehicles, obstacles and pedestrians). Then, the motion state of road vehicles and obstacles under surveillance is short-term predicted with regards to the Kalman Filter (KF), and the crossed vehicle trajectory will be detected half a second early. Furthermore, vehicular distance are predicted and compared with the estimated safe threshold, and the traffic crash risk can be evaluated. Finally, a group of simulation tests are conducted and the results indicate that the proposed system and method are of effective for detecting vehicle crash risk and providing driver accurate warning.