How do drivers negotiate horizontal ramp curves in system interchanges in the Netherlands?

Abstract

On interchanges there are higher probabilities of risky situations compared to uninterrupted motorway sections due to increased speed variability and higher frequency of lane-changes. In this study, we focus on understanding and modelling drivers' longitudinal speed behavior when negotiating horizontal ramp curves in interchanges in the Netherlands. For this purpose, detailed trajectory data of free-moving vehicles on 29 different curves from 6 different interchanges were collected from video images taken from a hovering helicopter. Only free-moving vehicles were chosen in order to understand how the road geometric design affects the (unhindered) driving speeds. The results of the speed profiles analysis show that for each connection, the speed profiles follow certain patterns, despite the large heterogeneity among drivers. These speed patterns were found to be significantly affected by the distance along a connection, the design characteristics of a connection, vehicle type, and drivers' heterogeneity. The impact of the distance along the connection on the speed was found to be significant and non-linear. This indicates that drivers do not maintain constant speeds, but adapt it along the connections. These models, which describe drivers' speed behavior and adaptation along different connections, are useful for improving current speed behavior models used in different microscopic simulation packages, and provide designers with a tool to estimate the speeds during the design process. The insights from this study, and the identified models, are also useful for enhancing the acceptability of automated vehicles' longitudinal behavior by adapting it to human like behavior.

Modelling overtaking strategy and lateral distance in car-to-cyclist overtaking on rural roads: A driving simulator experiment

“…Random effects allow the residuals associated with the longitudinal measures on the same unit of analysis to be correlated, thus taking into account the clustering effect. Linear mixed-effect models have been used in previous studies to model driver overtaking behaviour (Farah, 2013), driver speed behaviour on curves (Farah, Daamen, & Hoogendoorn, 2018), and the modelling choices of control transitions in automated driving (Varotto, Farah, Toledo, van Arem, & Hoogendoorn, 2017). The models developed in this study were estimated using the 'lme4' package (Bates et al, 2014) and validated using the 'rpart' package (Therneau, Atkinson, Ripley, & Ripley, 2018) in the R statistical program (Team, 2013).…”

Safety evaluation of freeway interchange merging areas based on driver workload theory

“…(LF/HF) m was the mode value of (LF/HF) i when the driver maneuvered on a flat straight segment of freeway, without overtaking, changing lane, speaking and answering phone calls, and so on. The K value in real time was calculated using equation (1), and part of the processed data in a merging area is shown in Table 3.…”

“…The LF/HF ratio measures the cardiac balance of the automatic nervous system and basically represents the sympathetic nervous tension and can further indicate the driving workload. 23 To characterize the impact of driving speed on HRV, driver workload can be quantified by equation (1), with the safety levels as shown in Table 1. 24,25 (LF/ HF) m is the HRV baseline for each driver and it is the mode value of (LF/HF) i in real time when maneuvering on a flat straight segment of freeway.…”

“…Freeway interchanges are high-risk areas because of their intricate geometric configuration, the demand for lane-changing, and variable driving speeds. 1 Previous studies have found that the rate of crashes per kilometer is much higher in interchange areas than in other freeway segments. 2 According to crash statistics from 2014 to 2016 recorded by the transportation administration department in Jiangsu Province in East China and collected and analyzed in this study, crashes over the three years across the whole province numbered 44,256, with 971 deaths and 4819 injuries.…”

Identifying and correcting the errors of vehicle trajectories from roadside millimetre‐wave radars

“…The potential value of full sample vehicle trajectories has been demonstrated in various applications, including driving behaviour analysis [10][11][12], road safety state estimation [13,14], identification and prediction of high-risk driving behaviour [15,16] and crashes [17][18][19], traffic monitoring [20,21], and active traffic safety management [22]. The first two functions are usually performed offline and the rest are often applied online.…”