Reliable communication stack for flexible probe vehicle data collection in vehicular ad hoc networks

Traffic congestions caused by high vehicular densities are an ever increasing problem for both personal and professional transportation, resulting in significant losses each year. While expanding the road infrastructure often offers a short term solution, more intelligent approaches are necessary to
equally distribute the traffic demand throughout the available infrastructure. To enable intelligent road infrastructure traffic management, detailed knowledge about both current and the historical states are necessary. Equipping vehicles with processing and communication devices enables them to sense their own state and that of their environment through existing built-in sensors as they travel through the road infrastructure. This makes the vehicles ideal candidates for generating both traffic information that is relevant for other drivers as well as any other type of information generated by sensors carried by vehicles given a possibility to support applications that go beyond traffic monitoring.
A key challenge in the collection of probe vehicle data, i.e., sensor data generated by vehicles, is to balance the resources used for data collection and benefits from data usage. Given variable application requirements, large amounts of vehicles and their increasing number of sensors, it is infeasible to collect all sensor information everywhere and at all times due to resource constraints.
Rather, the amount of generated and collected sensor data should be limited to only what is strictly necessary, by applying efficient data collection strategies.
The purpose of this work is to analyze and improve the probe vehicle data collection from a management and communication point of view, in the context of an 802.11p based, challenged communications infrastructure. We analyze the possibilities of using a distributed set of access points, so called road-side units, to facilitate the collection of probe vehicle data from the vehicles and propose
methodologies that mitigate the associated challenges. Specifically, we define how management of probe vehicle data can be realized in such an environment, allowing application to specify when and what sensor data they need. From the communication point of view the main thesis contributions are three-fold;
1) We define and evaluate an approach that enables the distribution of a communication session over multiple road-side units when the communication requirements exceed the communication resource a single access point can provide for delay tolerant applications.
2) We improve the information exchange between road-side units and vehicles by identifying communication characteristics of the road-side unit and use them to determine the optimal location at which the information exchange should occur.
3) We extend the coverage range of the road-side units through vehicle to
vehicle communication by modifying an existing routing algorithm, improving both delivery rate and communication overhead.
Applying the proposed methodologies on the collection of probe data provides applications with a set of tools that can be used to realize their requirements. Management of what data is collected frees up resource for a wider range of applications as unnecessary data collection can be switch off when not needed, while at the same time making it possible to increase resolution if needed. Using
other vehicles as forwarders can reduce the collection delay from minutes to seconds, if the network topology allows it, and scheduling the communication at the optimal distance to the road-side unit reduces communication overhead and distributes the resource consumption over time.