The state-of-the-art advancements in wireless communication and pervasive internet capabilities have led to significant breakthroughs in intelligent transportation systems where connectivity, autonomous driving and infotainment enhance the driving experience. The Vehicle ad hoc Networks has emerged as a distinguished branch of wireless communication; that pertains categorically to transportation systems. In the recent years, the evolutionary arc, depth and scope of internet utilities in the transportation systems, particularly for on-road vehicles, have been applied to solve some of the security and road-safety issues. The road transportation system with the help of VANET is being evolved into a safer and efficient architecture. This paper introduces a conceptual design of VANET based novel driver-behavior detection system that utilizes wireless sensors and the concept of context-awareness, marking a vital step towards improving road safety. The driver-behavior is defined as an uncertain context characterized by constant interaction between the driver, the vehicle and the environment. This paper introduces novel 3-tier architecture for real-time driver-behavior detection that uses Swarm Intelligence to reason about the contextual uncertainty and to deduce the driver behavior. The system is designed to detect five styles of driving behavior viz.: drunk, wayward, rash, fatigued/drowsy and acceptable driving and sends out an alert if the driver is detected in an undesired state while driving. The system is proposed to be implemented in Network Simulator-3. The objective of this paper is to contribute towards developing suitable safety measures for road transport where drivers, smart vehicles and road-side infrastructure work in collaboration using the available context information.