Analysing road user behaviour at roadwork zones using large-scale trajectory data

Roadwork zones are essential for maintaining and expanding the road network. However, during their active phases, they often lead to reduced traffic efficiency and increased safety risks for both road users and workers. To mitigate these effects, stationary roadworks typically lasting several hours employ measures such as lane closures and variable message signs (VMS) to influence driver behaviour and improve safety and efficiency. While merging behaviour and speed compliance have been studied at roadwork sites, few studies have used large-scale trajectory data to analyse detailed driver behaviour. This study addresses that gap by analysing trajectory data from a stationary roadwork zone on the E20 highway in Sweden, focusing on merging hotspots and undesired road user behaviour. Through the analysis of trajectory data, we aim to identify patterns that occur among multiple road users, as well as recurrent undesired behaviours among road users. Such insights provide an evidence-based foundation for refining roadwork design to influence behavioural patterns and mitigate associated risks. Results show that lane-end signs strongly influence merging locations, creating distinct hotspots where drivers tend to change lane near each sign. Drivers merging closer to the lane closure tend to travel at higher speeds, often exceeding posted limits. A small but notable share of drivers misinterpret signage near the on-ramp, leading to hazardous manoeuvrers such as unnecessary entry into the acceleration lane. These findings highlight the importance of careful sign placement and roadwork design to promote safe and efficient driver behaviour.