The vehicle trajectories analysis on dangerous bends is an important task to improve road safety. This paper propose a new methodology to predict failure trajectories of light vehicles in curve driving. It consists to use a stochastic modelling and reliability analysis in order to estimate the failure probability of vehicle trajectories. Firstly, we build probabilistic models able to describe real trajectories in a given bend. The models are transforms of scalar normalized second order stochastic processes which are stationary, ergodic and non-Gaussian. The process is characterized by its probability density function and its power spectral density estimated starting from the experimental trajectories. The probability density is approximated by using a development on the basis of Hermite polynomials. The second part is devoted to apply a reliability strategy intended to associate a risk level to each class of trajectories. Based on the joint use of probabilistic methods for modelling uncertainties, reliability analysis for assessing risk levels and statistics for classifying the trajectories, this approach provides a realistic answer to the tackled problem. Experiments show the relevance and effectiveness of this method.
The vehicles real trajectories analysis on dangerous zones is an important task to improve the road safety. The objective of this study is to provide tools for driving behaviour identification with the associated risk as regards of handling loss. This study aims to take into account the infrastructure, driver and the vehicle interactions, which are useful to evaluate this risk accurately. We propose in this paper a vehicles trajectories analysis in bend within a suitable Support Vector Machine (SVM) algorithm framework. At first, we will be interested on vehicle trajectory definition and exper imental data acquisition. Then, we will make an experimental trajectories classification in order to determine several classes of trajectories. Afterwards, we will make a vehicle trajectories stability analysis in order to identify safe and unsafe fields of the observed trajectories. Lastly, one will use machine learning methods to predict failure trajectories.978-1-4244-3834-1/09/$25.00
The new approaches based on the credal partition commit objects; not only to singleton clusters but also to meta-clusters; with different masses of beliefs. Among these approaches we find the Credal C-means (CCM). In this article, we introduce a new clustering algorithm called kernel Credal C-means (KCCM), which is a kernel version of CCM algorithm. It is based on the beliefs functions. A second contribution of this paper is an improved CCM algorithm using a non-Euclidean metric unlike CCM which is based on Euclidean distance. In order to show the effectiveness of the proposed method, artificial and real data (vehicle trajectory in a bend) are tested at the end of this paper. A comparison is made to evaluate the results obtained with some methods as Evidential Cmeans (ECM) and CCM.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.