Piotr Szczurek

Learning the relevance of parking information in VANETs

The use of Vehicular Ad-Hoc Network (VANET) has been applied to many applications involving information dissemination. Many of such applications are limited by the communication limitations of a VANET, such as limited transmission range and bandwidth. This imposes a necessity for evaluating the relevance of information. This paper proposes the use of machine learning for finding relevance of information for a parking information dissemination system. The proposed method uses the learned relevance for aiding vehicles in decision making by finding the probability that a given parking location will be available at the time of arrival. The method was evaluated through simulations and the results show that the proposed method is successful at learning the relevance of parking reports, which resulted in lower parking discovery times for vehicles.

Machine learning approach to report prioritization with an application to travel time dissemination

This paper looks at the problem of data prioritization, commonly found in mobile ad-hoc networks. The proposed general solution uses a machine learning approach in order to learn the relevance value of reports, which represent sensed data. The general solution is then applied to a travel time dissemination application. Through the use of offline learning, the paper analyzes the feasibility of the proposed approach and compares the accuracy performance of several common machine learning algorithms. The results show that not all machine learning algorithms may be used for prioritization and that the use of the logistic regression algorithm is particularly suited for the problem. The learned logistic regression model is then used in a simulated VANET environment. The results of the simulations show that it is better at prioritizing reports in terms of their usefulness in aiding vehicles to choose the shortest travel time paths.

Estimating Relevance for the Emergency Electronic Brake Light Application

In this paper, we compare two methods of estimating relevance for the emergency electronic brake light application: One method uses an analytically derived formula based on the minimum safety gap that is required to avoid a collision, whereas the other method uses a machine learning approach. The application works by disseminating reports about vehicles that perform emergency deceleration in an effort to warn drivers about the need to perform emergency braking. Vehicles that receive such reports have to decide on whether the information contained in the report is relevant to the driver and warn the driver if that is the case. Common ways of determining relevance are based on the lane or direction information, but using only these attributes can lead to many false warnings, which can desensitize the driver. Desensitized drivers may ignore warnings or completely turn off the system, thus eliminating any safety benefits of the application. We show that the machine learning method, compared with the analytically derived formula, can significantly reduce the number of false warnings by learning from the actions that drivers take after receiving a report. The methods were compared using simulated experiments with a range of traffic and communication parameters.

Intelligent transportation systems: When is safety information relevant?

In this paper, we compare two methods of estimating relevance for the emergency electronic brake light application. One uses an analytically derived formula based on the minimal safety gap required to avoid a collision. The other method uses a machine learning approach. The application works by disseminating reports about vehicles that are performing emergency deceleration in effort to warn drivers about the need to perform emergency braking. Vehicles receiving such reports have to decide whether the information contained in the report is relevant to the driver, and warn the driver if that is the case. Common ways to determine relevance are based on the lane or direction information, but using only these attributes can still lead to many false warnings, which can desensitize the driver. Desensitized drivers may ignore warnings or turn off the system completely, thus eliminating any safety benefits of the application. We show that the machine learning method, in comparison to the analytically derived formula, is able to significantly reduce the number of false warnings by learning from the actions drivers take after receiving a report. The methods were compared using simulated experiments with a range of traffic and communication parameters.

A methodology for the development of novel VANET safety applications

We present a methodology for the development of passive ITS safety applications that aim to disseminate reports about dangerous events on the road. Examples of such applications include the emergency electronic brake light or the highway merge warning. A major issue with such applications is the decision of when a warning should be shown. Since the recipient vehicle may be far away from where the dangerous event occurred, a large number of false warnings may be shown to the drivers. This leads to driver desensitization which may reduce the safety benefits. While previous research has provided a way of handling false warnings by estimating the relevance of the reports, these methods do not take into all the important factors and are not easily adaptable for novel applications. In this paper, we propose a simulation platform for developing and evaluating relevance estimators for passive ITS safety applications that can be utilized for developing novel applications. The paper provides examples of the effectiveness of this platform on three previously proposed applications.

A platform for the development and evaluation of passive safety applications

In this paper, we present a platform for aiding in the development and evaluation of novel ITS passive safety applications. Such applications work by having vehicles detect certain events that may be dangerous to other vehicles and disseminating reports about these events using wireless communication. A vehicle receiving the report about the event can then be warned. However, a large number of false warnings will lead to driver desensitization, which will reduce the safety benefit. To overcome this issue, a relevance estimator that will determine for which reports a warning will be given has to be devised for each new application. Our platform allows for an easy, fast method of developing these estimators and evaluating them in simulations. We demonstrated the feasibility of this approach with three example applications.

A Traffic Analysis Perspective on Communication in the Brain

In this short paper, we report on an approach to datamine the brain from a novel perspective, namely traffic analysis. Our datamining approach considers the brain regions and the tracts that connect them as a road network, and the signals traveling between them as vehicles. We analyze travel patterns by a process called traffic assignment. The results are unexpected in the sense that the movement of signals in the brain seems to follow some global optimization patterns as opposed to the anarchical system that would be favored by evolution.

Observe-Driver-and-Learn Platform for Relevance Estimation in Safety Warning Applications from Vehicular Ad Hoc Network

Safety warning applications from a vehicular ad hoc network aim to disseminate alerts about dangerous events on the road with wireless communication technology and, when necessary, warn drivers receiving such alerts. Examples include the emergency electronic brake light or the highway merge warning. A major issue with such applications is false warnings, which lessen any safety benefits the applications provide. A high number of false warnings will lead to driver desensitization and reduce any potential safety benefits. Therefore any received alert has to be evaluated in terms of its relevance for the given vehicle. However, the relevance depends on a combination of many factors and is specific to a given application, so defining an estimator is difficult. A machine-learning method based on the principle of observe-driver-and-learn is proposed for finding relevance estimators. This method is evaluated for its effectiveness with three safety applications: electronic emergency brake lights, the highway merge warning, and the control loss warning.