Robust statistics is a branch of statistics which includes statistical methods capable of dealing adequately with the presence of outliers. In this paper, we propose an anomaly detection method that combines a feature selection algorithm and an outlier detection method, which makes extensive use of robust statistics. Feature selection is based on a mutual information metric for which we have developed a robust estimator; it also includes a novel and automatic procedure for determining the number of relevant features. Outlier detection is based on robust Principal Component Analysis (PCA) which, opposite to classical PCA, is not sensitive to outliers and precludes the necessity of training using a reliably labeled dataset, a strong advantage from the operational point of view. To evaluate our method we designed a network scenario capable of producing a perfect ground-truth under real (but controlled) traffic conditions. Results show the significant improvements of our method over the corresponding classical ones. Moreover, despite being a largely overlooked issue in the context of anomaly detection, feature selection is found to be an important preprocessing step, allowing adaption to different network conditions and inducing significant performance gains.

A queuing system resulting from a signalized intersection regulated by pretimed control in a network of urban traffic is considered. Modeling the queue length and the delay of vehicles is crucial to evaluate the performance of intersections equipped with traffic signals. Air quality and rational use of energy also depend on an efficient management of the intersections. These traffic systems have the specificity that the server (green signal) is deactivated (red signal) during a fixed period of time. In the present work, an M/D/1 queue with a server that occasionally takes vacations is analyzed. The mean delays of vehicles and the mean queue length are computed and compared with those obtained by using a detailed simulation model in a case study. We find that, in general, the mean delays of vehicles given by the proposed queuing model provide a good approximation, but the model can provide slightly smaller values than those obtained in the simulation model for large traffic flows. This result is of interest for traffic engineers, as the approaches one can find in the literature for large signalized urban traffic flows are subject to criticism.