Carchat: A vehicular wireless ad-hoc network data dissemination protocol

Madalina Baker, Purdue University

Abstract

In vehicular wireless ad-hoc networks, it is often useful that certain information be made available to all vehicles within a reasonable amount of time. Since traffic can be stalled or disrupted in urban areas due to many possible occurrences, having knowledge of these events ahead of time can help in deciding to follow an alternate, unobstructed route. However, the motorists that could use this information are moving targets, and their unpredictable mobility makes any data dissemination a difficult goal to achieve. This type of data is already being collected in many cities and it is communicated to motorists through dedicated AM radio stations and warning display panels on highways. However, the dissemination of this information is strictly limited by the infrastructure in place and individual motorists have to pursue each data item individually, making it a cumbersome and potentially distracting process. Also, since there are large coverage gaps in the infrastructure placement, the information available is sparse and incomplete. Current data dissemination algorithms which are designed for static or low mobility ad-hoc networks become very ineffective in a mixed network of static and mobile nodes (i.e. static infrastructure and mobile vehicles). Also, there are dissemination algorithms that make use of the mobility of certain nodes to ferry information from a source to a certain destination. However, these algorithms generally involve propagating information from a single source to a single destination, which does not disseminate the data to all of the nodes in the network, but rather only transmits the data to intermediate nodes that serve as ferries. The goal of the CarChat protocol is to explore the mobility of vehicles in order to disseminate useful data farther than the reach of the existing infrastructure. Every vehicle would be equipped with a radio transceiver that can communicate with fixed infrastructure transmitters as well as with radios installed in other vehicles in the road. Each vehicle would then collect new information while in range of the infrastructure transmitters, as well as exchange most up-to date information with vehicles it encounters outside of this range. It would estimate the mobility of other vehicles to determine which would be good candidates to engage in data transfer. We test the efficiency of this protocol using testbed experiments as well as simulations, in which the vehicles exhibit three different driving behaviors - high speed highway driving, moderate speed rural driving, and congested urban driving.

Degree

M.S.E.C.E.

Advisors

Bagchi, Purdue University.

Subject Area

Computer Engineering

COinS