原文传递 State-of-the-art of micro-simulation models
题名: State-of-the-art of micro-simulation models
责任者: H Gibson
关键词: not available
学科分类: 交通工程
摘要: This paper reviews microsimulation techniques with a view to understanding how they could be enhanced and used to model Intelligent Transport Systems (ITS). Microsimulation tools model individual vehicles using ‘bottom up’ rules to determine each vehicle’s speed, acceleration, lane changing and so forth, which on aggregate provide the speed and flow on the highway. Microsimulation tools are able to adapt to new influences affecting vehicle behaviour; their downside is that many effects on the highway are complex and calibration of the tools is a problem. An accurate microsimulation model is vital for predicting how a new scheme or technology will perform, and could be used for predicting congestion in real time, thus allowing network operators to mitigate through speed control and route guidance. Microsimulation models are already widely used in assessing impacts on emissions, noise and fuel consumption. This report covers the range of techniques used to build the microsimulation engines that lie at the heart of the microsimulation tools. The importance of accurately calibrating the microsimulation is highlighted. The three market leading tools in Europe (AIMSUN, VISSIM and PARAMICS) are assessed in detail alongside SISTM, a product developed by TRL. In addition some 49 other tools (mostly developed for research purposes) are compared. This report recommends that significant improvements can be made to most microscopic modelling projects through enhanced calibration using CCTV, MIDAS loops and other highways data. Such data would also facilitate improvements to lane changing algorithms, so that lane changing characteristics and use of lanes are modelled better. The combined effect would be more reliable prediction of the impact of highway schemes.
出版机构: Transport Research Laboratory
提交日期: 8 January 2013
总页数: 26
报告类型: 科技报告
资源类型: 科技(咨询、行业)报告
初始创建时间: 8 January 2013
检索历史
应用推荐