Analyzing the Impact of Wireless Multi-Hop Networking On Vehicular Safety

One of the core challenges of Intelligent Transportation System is the dissemination of timely and accurate vehicle information (e.g. speed, position) to geographically large distances without compromising data supply rates from immediate neighbors. This feature is critical for the design of vehicle safety and navigation applications. Single hop broadcasting is often inadequate to ensure vehicle safety when the platoon size is arbitrarily large due to its upper bound on rate and range of wireless message transmission. Existing wireless multi-hop protocols do not ensure reliable message delivery while avoiding network congestion in the shared channel. In this thesis, we make two separate but related investigations to address this challenge - (1) Analyze the impact of distance sensitive multi-hop broadcasting in realistic traffic network (2) Analyze the impact of wireless multi-hop network in vehicle safety. For investigating the first part, we used VCAST, a distance sensitive information propagation technique, in which information is forwarded at a rate that decreases linearly with distance from the source. VCAST is evaluated by using extensive simulations in ns-3, a discrete event simulator for wireless and mobile ad-hoc networks, under different density, source broadcast rates and communication range. To simulate realistic traffic movement, we used 2d grids of different sizes and used both uniform and non-uniform mobility. The results show that VCAST is scalable for - large number of vehicles and large source broadcast rates. It is further shown that successful scaling is achieved by reduced number of vehicle records transmitted per second per vehicle for varying network sizes and varying source broadcast rates. Vehicle safety messages for VCAST are piggy backed on heart beat messages and does not require any modifications to the existing vehicular communication standards. For investigating the second part, we implemented a realistic car following model and used string stability analysis as a metric for measuring vehicle safety. The basic idea is to exploit the small network propagation time in disseminating safety messages over large distances, instead of relying on just the predecessor vehicle\u27s state. This enables distant vehicles in a traffic stream to plan well in advance against rear end collisions which could lead to string instability. We also proposed one such proactive method of planning - and that is by controlling the headway time. Through extensive simulations, we obtained results for vehicle safety when some incident is detected abruptly on its course. The results show that proactive planning using multi-hop network makes the entire platoon string stable in the presence of emergency road incidents