Sophie Midenet

Exposure to lateral collision in signalized intersections with protected left turn under different traffic control strategies

This paper proposes an original definition of the exposure to lateral collision in signalized intersections and discusses the results of a real world experiment. This exposure is defined as the duration of situations where the stream that is given the right-of-way goes through the conflict zone while road users are waiting in the cross-traffic approach. This measure, obtained from video sensors, makes it possible to compare different operating conditions such as different traffic signal strategies. The data from a real world experiment is used, where the adaptive real-time strategy CRONOS (ContRol Of Networks by Optimization of Switchovers) and a time-plan strategy with vehicle-actuated ranges alternately controlled an isolated intersection near Paris. Hourly samples with similar traffic volumes are compared and the exposure to lateral collision is different in various areas of the intersection and various traffic conditions for the two strategies. The total exposure under peak hour traffic conditions drops by roughly 5 min/h with the CRONOS strategy compared to the time-plan strategy, which occurs mostly on entry streams. The results are analyzed through the decomposition of cycles in phase sequences and recommendations are made for traffic control strategies.

Creating ensemble classifiers through order and incremental data selection in a stream

This paper presents an original time-sensitive traffic management application for road safety diagnosis in signalized intersections. Such applications require to deal with data streams that may be subject to concept drift over various time scales. The method for road safety analysis relies on the estimation of severity indicators for vehicle interactions based on complex and noisy spatial occupancy information. An expert provides imprecise labels based on video recordings of the traffic scenes. In order to improve the performance—overall and for each class—and the stability of learning in a stream, this paper presents new ensemble methods based on incremental algorithms that rely on their sensitivity to the processing order of instances. Different data selection criteria, many used in active learning methods, are studied in a comprehensive experimental evaluation, including benchmark datasets from the UCI machine learning repository and the prediction of severity indicators. The best performance is obtained with a criterion that selects instances which are misclassified by the current hypothesis. The proposed ensemble methods using this criterion and AdaBoost have similar principles and performance, while the proposed methods have a smaller computational training cost.