MODELLING MOTORWAY DRIVING BEHAVIOUR USING MICROSCOPIC VEHICLE TRAJECTORY DATA

摘要

The paper presents a micro-simulation model of motorway driving using vehicle trajectory data to define key behaviours such as longitudinal and lateral movements. Micro-simulation plays an important role in transport research. In particular it can be used to appraise new transport schemes, help predict short/medium term traffic conditions in real time, and provide a realistic driver behaviour model for implementation within a driving simulator. In the past microscopic model developers have tended to build models 'top-down' due to the cost of acquiring suitable microscopic field data. A top-down approach has a principal aim of creating microscopic models that represent aggregated macroscopic field data correctly, with less attention to the behaviour at the level of individual drivers. The aim of the project described in this paper was to build a microscopic model from the 'bottom-up'. A bottom-up approach uses microscopic field data (namely vehicle trajectory data) to build and calibrate microscopic models, and has a principal aim to replicate the behaviour of individual drivers. It is anticipated that if the microscopic behaviour is modelled with sufficient accuracy then the aggregated macroscopic statistics (e.g. journey times, capacities, queue lengths, etc.) will also be correct, and greater confidence can be given to the model predictions.