Early detection of dangerous events on the road using distributed data fusion

Abstract

Intelligent transport systems are a fast developing area of research with great impact on everyday life. One of the main ideas in this area is to use all possible information, coming from vehicles and the infrastructure, in order to make the system "smarter" and avoid unwanted situations -collisions, accidents, bottlenecks... Sources of data are sometimes unreliable and may lead vehicles or the whole system to wrong conclusions and adaptations. We are presenting the application of a distributed data fusion algorithm to detect dangerous events on the road. This application is adapted to detect the possibility of encountering icy roads based on readings from wireless temperature sensors. It takes into account data not only directly obtained from sensors but also from the neighborhood of each element in the system. This way, we obtain a more robust solution that is flexible to the unreliabilities of the sources of data. We demonstrate the possibility to deduce proper results from unreliable data. The algorithm is tested in emulation, using data from a real testbed, to show the usefulness and correctness of our approach.

Detecting Road Events Using Distributed Data Fusion: Experimental Evaluation for the Icy Roads Case

“…This similarity between the on-the-road and the in-lab experiments confirm our previous results with one vehicle in [8]. In the following, we use network emulation to study a threevehicle scenario.…”

“…In this paper, we present a generic method to handle multiple data sources in unreliable vehicular networks, that can be used to design applications in the context of intelligent transportation systems for advance detection of dangerous events on the road. This work completes our previous paper [8]. It presents results from road experiments using sensors, road-side units (RSU) and vehicles equipped with on-board units (OBU) [9].…”