An observational study of understanding the factors influencing merging behaviour in work zones
Sajani Siriwardene, Muhammad Arif Ashraf, Ashim Kumar Debnath
Abstract
• Driver merging behavior observed at work zones using video recording. • Merging behavior differed between slow lane and fast lane closure setups. • Most vehicles merge late in slow lane closure (SLC) and early in fast lane closure (FLC) • Heavy vehicles in SLC merge at higher speed and lower acceleration than light vehicles. • Vehicles in FLC setup have higher speed variation, except just before the taper area. Merging due to lane closure in work zones is an everyday driving activity that is characteristically different from merging in non-work zone sections. Despite many studies analysing merging behaviour in lane change situations, limited efforts explicitly aimed to understand work zone merging processes, particularly using real-world observational data and to distinguish behavioural patterns for different types of lane closure configurations. To fill these important gaps, this paper analyses merging behaviour in work zones featuring slow lane closure (SLC) and fast lane closure (FLC) using observational data from work zones in Australia. Merging behaviour in terms of merging locations, merging speed, and acceleration/deceleration rate during merging were examined using regression models. The results showed that work zone merging behaviour varied significantly among the SLC and FLC types of lane closure. In an SLC setup, most vehicles merged closer to the taper end, whereas most vehicles in an FLC setup merged well ahead of the taper start area. Traffic and vehicle characteristics, including vehicle type, type of vehicle in front, gap to the target lane lag vehicle, traffic volume, proportion of heavy vehicles, and order of a merging vehicle in a platoon, influenced merging behaviour across both types of sites. Traffic simulation and design of future work zone traffic controls should consider the insights developed on different merging behaviour and their influential factors in SLC and FLC setups.