Lane Changing and Lane Utilisation Behaviour in the UK: Empirical Study

Lane Changing and Lane Utilisation Behaviour in the UK: Empirical Study


Hamid Athab Al-Jameel, University of Kufa


This paper focuses on finding the relationships between the flow rate of the road under study and frequency of the lane change and lane utilization.


Discernment of driver behaviour is the key to understanding dynamic driver behaviour for lane changing (LC) and lane utilization (LU). The LC and LU parameters are critically important for the calibration and validation of a traffic simulation model and therefore, this study focuses on LC and LU behaviour. Field data has been collected from examples of three types typical motorway cross-sections: two-lane (M602), three-lane(M60) and four lane(M60). All are located in the City of Manchester in the UK. Video cameras have been installed at suitable vantage points in order to observe traffic in these sections. The data analysis for these locations indicated that there is a relationship between volume of flow and the frequency of LC. As the flow increases, the number of instances of LC increases up to a certain flow but then declines if the traffic flow increases further. The value of flow for maximum LC depends on the section, whether two lane, three lane or four lane. The LU behaviour has also been investigated for each type of cross-sections under study. A number of empirical equations have been derived from this study which will be used in the calibration and validation of a simulation model for subsequent stage of this project. The main conclusions from this study could be summarised as: Firstly, It was found from field data that relationships can be established between the FLC and the flow rate for two and four lane sections. These relationships will be used for calibration and validation for any simulation model. Secondly, Various lane utilisation relationships were investigated for different types of normal sections (two, three, and four) under varying levels of flow and other conditions. These relations will be used as initial inputs to distribute traffic across different lanes in any simulation model. Thirdly, The LU has also been investigated for HGVs. This is also an important factor to allow accurate calibration or validation for a simulated model. Finally, Using the MIDAS data has provided sufficient information about the flow regime on normal section which will be used in investigating the LU factor.


Association for European Transport