Aliaksei Laureshyn

Evaluation of traffic safety, based on micro-level behavioural data: Theoretical framework and first implementation

A traffic encounter between individual road users is a process of continuous interplay over time and space and may be seen as an elementary event with the potential to develop into an accident. This paper proposes a framework for organising all traffic encounters into a severity hierarchy based on some operational severity measure. A severity hierarchy provides a description of the safety situation and trade-off between safety and efficiency in the traffic system. As a first approach to study the encounter process, a set of indicators is proposed to describe an encounter. These indicators allow for a continuous description even if the relationship between the road users changes during the process (e.g., when they are on a collision course or leave it). Automated video analysis is suggested as a tool that will allow data collection for validation of the proposed theories.

Cyclists in roundabouts -- Different design solutions.

Whether the safest roundabout design for cyclists is to separate cycle crossings or integrate cyclists with motorists is an extensively discussed issue. Studies using accident statistics indicate that a separated cycle crossing is the safest for high motor vehicle volumes. However, the results have not been satisfyingly explained. This article combines quantitative and qualitative methods in traffic conflict, interaction and behavioural studies to find out how interactions and conflicts differ between the two roundabout designs. Automated video detection is used as one of the methods and its performance is evaluated. The integrated roundabout turns out to be more complex with a higher number of serious conflicts and interaction types. The most dangerous situations in the integrated roundabout seem to come about when a motorist enters the roundabout while a cyclist is circulating and when they are both circulating in parallel and the motorist exits. The yielding rules are more ambiguous in the separated roundabout, contributing to a lower yielding rate to cyclists and a lower trust in the other road users willingness to yield. Situations in the separated roundabout with the lowest yielding rate to cyclists occur when the motorist exits the roundabout at the same time as cyclists are riding in the circulating direction and hence coming from the right. However, most of the accidents in separated roundabouts occur while cyclists are riding against the circulating direction, both when motorists enter and exit the roundabouts.

In search of the severity dimension of traffic events: Extended Delta-V as a traffic conflict indicator

Most existing traffic conflict indicators do not sufficiently take into account the severity of the injuries resulting from a collision had it occurred. Thus far, most of the indicators that have been developed express the severity of a traffic encounter as their proximity to a collision in terms of time or space. This paper presents the theoretical framework and the first implementation of Extended Delta-V as a measure of traffic conflict severity in site-based observations. It is derived from the concept of Delta-V as it is applied in crash reconstructions, which refers to the change of velocity experienced by a road user during a crash. The concept of Delta-V is recognised as an important predictor of crash outcome severity. The paper explains how the measure is operationalised within the context of traffic conflict observations. The Extended Delta-V traffic conflict measure integrates the proximity to a crash as well as the outcome severity in the event a crash would have taken place, which are both important dimensions in defining the severity of a traffic event. The results from a case study are presented in which a number of traffic conflict indicators are calculated for interactions between left turning vehicles and vehicles driving straight through a signalised intersection. The results suggest that the Extended Delta-V indicator seems to perform well at selecting the most severe traffic events. The paper discusses how the indicator overcomes a number of limitations of traditional measures of conflict severity. While this is a promising first step towards operationalising an improved measure of traffic conflict severity, additional research is needed to further develop and validate the indicator.

From Speed Profile Data to Analysis of Behaviour: Classification by Pattern Recognition Techniques

Classification of speed profiles is necessary to allow interpretation of automatic speed measurements in terms of road user behavior. Aggregation without considering variation in individual profile shapes easily leads to aggregation bias, while classification based on exogenous criteria runs the risk of losing important information on behavioral (co-) variation. In this paper, the authors test how 3 pattern recognition techniques (cluster analysis, supervised learning, and dimension reduction) can be applied to automatically classify the shapes of speed profiles of individual vehicles into interpretable types, with a minimum of a priori assumptions. The data for the tests is obtained from an automated video analysis system and the results of automated classification are compared to the classification by a human observer done from the video. Normalization of the speed profiles to a constant number of data points with the same spatial reference allows them to be treated as multidimensional vectors. The k-means clustering algorithm groups the vectors (profiles) based on their proximity in multidimensional space. The results are satisfactory, but still the least successful among the tested techniques. Supervised learning (nearest neighbor algorithm tested) uses a training dataset produced beforehand to assign a profile to a specific group. Manual selection of the profiles for the training dataset allows better control of the output results and the classification results are the most successful in the tests. Dimension reduction techniques decrease the amount of data representing each profile by extracting the most typical “features”, which allows for better data visualization and simplifies the classification procedures afterwards. The singular value decomposition used in the test performs satisfactorily. The general conclusion is that pattern recognition techniques perform well in automated classification of speed profiles compared to classification by a human observer. However, there are no given rules on which technique will perform best.

Reduced Search Space for Rapid Bicycle Detection

This paper describes a solution to the application of rapid detection of bicycles in low resolution video. In particular, the application addressed is from video recorded in a live environment. The future aim from the results in this paper is to investigate a full year of video data. Hence, processing speed is of great concern. The proposed solution involves the use of an object detector and a search space reduction method based on prior knowledge regarding the application at hand. The method using prior knowledge utilizes random sample consensus, and additional statistical analysis on detection outputs, in order to define a reduced search space. It is experimentally shown that, in the application addressed, it is possible to reduce the full search space by 62% with the proposed methodology. This approach, which employs a full detector in combination with the design of a simple and fast model that can capture prior knowledge for a specific application, leads to a reduced search space and thereby a significantly improved processing speed. (Less)

A Search Space Strategy for Pedestrian Detection and Localization in World Coordinates.

The focus of this work is detecting pedestrians, captured in a surveillance setting, and locating them in world coordinates. Commonly adopted search strategies operate in the image plane to address the object detection problem with machine learning, for example using scale-space pyramid with the sliding windows methodology or object proposals. In contrast, here a new search space is presented, which exploits camera calibration information and geometric priors. The proposed search strategy will facilitate detectors to directly estimate pedestrian presence in world coordinates of interest. Results are demonstrated on real world outdoor collected data along a path in dim light conditions, with the goal to locate pedestrians in world coordinates. The proposed search strategy indicate a mean error under 20 cm, while image plane search methods, with additional processing adopted for localization, yielded around or above 30 cm in mean localization error. This while only observing 3-4% of patches required by the image plane searches at the same task. (Less)

How Accurately Can We Measure from Video? Practical Considerations and Enhancements of the Camera Calibration Procedure:

The accuracy of position measurements from videos depends greatly on the quality of the camera calibration model parameters. This paper investigates how such factors as camera height and the selection of calibration points affect the quality of the final calibration model. A series of controlled experiments were performed in traffic or similar-to-traffic environments, in which the accuracy of measurements from videos was compared with measurements taken with other tools. To enhance the calibration process, a multi-camera approach is suggested that utilizes the information about “common points” – points seen on several cameras but with unknown world coordinates. The performed tests showed that calibration quality can greatly benefit from this approach. The paper is addressed primarily to traffic researchers developing their own video-based tools for road user observations.

In search of surrogate safety indicators for vulnerable road users: a review of surrogate safety indicators

ABSTRACT Surrogate indicators are meant to be alternatives or complements of safety analyses based on accident records. These indicators are used to study critical traffic events that occur more frequently, making such incidents easier to analyse. This article provides an overview of existing surrogate indicators and specifically focuses on their merit for the analyses of vulnerable road users and the extent to which they have been validated by previous research. Each indicator is evaluated based on its ability to consider the collision risk, which can be further divided into the initial conditions of an event, the magnitude of any evasive action and the injury risk in any traffic event. The results show that various indicators and their combinations can reflect different aspects of any traffic event. However, no existing indicator seems to capture all aspects. Various studies have also focused on the validity of different indicators. However, due to the use of diverse approaches to validation, the large difference in how many locations were investigated and variations in the duration of observation at each location, it is difficult to compare and discuss the validity of the different surrogate safety indicators. Since no current indicator can properly reflect all the important aspects underlined in this article, the authors suggest that the choice of a suitable indicator in future surrogate safety studies should be made with considerations of the context-dependent suitability of the respective indicator.