196,142 research outputs found
Sensitivity analysis and how to choose parameters to calibrate
No full textThe MULTITUDE Project (Methods and tools for supporting the Use, caLibration and validaTIon of Traffic simUlations moDEls) is an Action (TU0903) supported by the EU COST office (European Cooperation in Science and Technology) and focuses on the issue of uncertainty in traffic simulation, and of calibration and validation as tools to manage it. It is driven by the concern that, although modelling is now widespread, we are unsure how much we can trust our results and conclusions. Such issues force into question the trustworthiness of the results, and indeed how well we are using them.
The project consists of 4 Working Groups (WGs) which hold short, focussed, meetings on topics of interest and propose work items on key issues. Additionally the project holds an annual meeting, as well as training schools, where the latest thinking can be passed on to young researchers and practitioners.
This report covers much of the technical work performed by Working Group 4 ‘Synthesis, dissemination and training’, and has been contributed to by:
- Costas Antoniou, NTUA, GR
- Jaume Barcelo, UPC, ES
- Mark Brackstone, IOMI, UK
- Hilmi Berk Celikoglu, ITU, TR
- Biagio Ciuffo, JRC, IT
- Vincenzo Punzo, JRC/UNINA, IT
- Pete Sykes, PS-TTRM, UK
- Tomer Toledo, Technion, IL
Peter Vortisch, KIT, DE
Peter Wagner, DLR, DE
This document assesses the current situation regarding guidelines for traffic simulation model calibration and validation worldwide, discusses the problems currently faced, and suggests potential ways in which they can be addressed, both directly, and indirectly through the development of the overall field of traffic simulation as a whole
The microscopic modelling of traffic flow: Weaknessess and potential developments
No full textIn: Workshop on Traffic and Granular Flow. Wolf, D.E., Schreckenberg, M. and Bachem, A. Eds. pp151-166. World Scientific, 1996
An instrumented vehicle for microscopic monitoring of driver behaviour
No full textA vehicle-based data collection system is explored. It is capable of gathering accurate data on the relative movements of adjacent vehicles, and hence driver behavior, in a traffic stream. The instrumented vehicle may support a variety of detection devices, for example a tracking laser rangefinder head, or a constantly scanning laser beam (a LIDAR). To relate the measurements to the position of the vehicles on the road, several sub-systems are devised to calculate corrections required by variations in steering, and vibrational movements caused by the road surface. Operational specifications of the vehicle are devised, in terms of road and traffic conditions required for optimum performance, and strategies for use.<br/
Validity of microscopic modelling of motorway traffic
No full textDevelopments have recently been made in increasing the quality of data available for the calibration of microscopic simulation models, and several techniques devised for high accuracy collection. Little consideration has been given however to the acceptability of the microscopic models themselves. This paper describes the broad features of car following models in current use, their data requirements and above all, examine the pertinence of current validation techniques to the processes under examination and their intended uses.<br/
A comparison of EU and US progress in the development of longitudinal (AVCSS)
No full textRecent years have seen the rapid development of advanced vehicle control and safety systems (AVCSS), dedicated to improving capacity and safety, particularly on multilane highways, by modifying driver behaviour, or introducing a degree of vehicle control. The extent and scale of these activities has resulted in many overlapping targets and approaches that have not been fully reconciled. This paper sets out to help develop a common understanding by: (1) reviewing the current state of the art of this subject in the European Union (EU) and the USA (2) assessing the potential benefits/dis-benefits of such systems and (3) assessing the barriers remaining before these systems may generally be implemented
Driver behaviour and traffic modelling . Are we looking at the right issues?
No full textAlthough much attention has been given to the simulation and modeling of driver behaviour, and comparison and testing of differing algorithms (such as car following) is now performed, there are several assumptions in use regarding micro-modeling that may not be correct. These could have important implications to our ability to model the impact of ITS (Intelligent Transport Systems), in particular, in-vehicle systems. In this paper we will examine four assumptions regarding car following models that may be in need of revision: 1) drivers adopt constant time headways and use 'safe' following distances, 2) there is a lack of data against which to undertake calibration/validation, 3) short time steps allow more realistic simulations of dynamics, and 4) that there are important' Chaotic patterns' in car followin
- …
