Communication requirements for crash avoidance

Safety applications are a driving force behind VANET deployment. Automobile manufacturers, government organizations, and consortia of the two have been investigating using VANETs for safety applications. Though VANETs are in large part designed for safety applications, researchers do not yet know the communication requirements of VANET safety messages. As a result, protocol designers have relied on generic network success metrics, such as packet delivery ratio, to evaluate their protocols. However, a more useful metric is the ability of currently proposed VANET schemes (e.g., for authentication, power control, etc.) to allow vehicles to receive safety messages and warn their drivers sufficiently in advance of an accident so that the driver can avoid the accident. Besides the basic safety message service, researchers have proposed other VANET mechanisms and services including mix zones [2, 5] and silent periods [18, 8, 17] to enhance vehicle privacy, intelligent transportation systems [24], and commercial applications [12]. However, these applications face a similar question: will a VANET be able to support these services and still achieve the safety goals for which the VANET was designed? Previous attempts at answering the above questions have been made using small test beds without any collisions and using vehicle kinematics and message reception probabilities. However, each of these approaches lack the realism (i.e., actual crashes) and scale that VANETs will have. In this paper, we present our results from simulating two vehicular safety applications. We simulated crash scenarios and determined the probability that vehicles could avoid the crashes. Additionally, we measured the communication requirements needed for those probabilities.