Human-like speed modeling for autonomous vehicles during car-following at intersections

Tawfeek, Mostafa H.

doi:10.1139/cjce-2020-0761

Cited by 3 publications

(1 citation statement)

References 46 publications

(28 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, researchers have studied three machine learning techniques to model drivers' speed in intersection scenarios during the vehicle's braking stage. The goal is to estimate a safe, comfortable, and human-like speed at the minimum distance for autonomous vehicles [37,38]. A microsimulation platform was created to evaluate the improved intersection car-following model and simulate the influenced saturation flow rate, based on an enhanced intelligent driver model [39].…”

Section: Related Workmentioning

confidence: 99%

Research on the Vehicle-Behavior Boundary of Intersection Traffic Based on Naturalistic Driving Data Study

Wu,

Ma,

Zhu

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

With the development and application of vehicle-infrastructure cooperative technology, the traffic regional safety related to intelligent connected vehicles (ICVs) has become the hotspot of the intelligent transportation system (ITS), and the integration of mixed autonomous and non-autonomous vehicles that are not cooperative in intersection areas has become a significant challenge due to the rapid advancement of autonomous vehicle technology. Autonomous vehicles in intersections with strong-structure and weak-rule characteristics pose a potential hazard in complex traffic situations. Studying the driving behavior of vehicles in intersections is of great significance due to the complex traffic environment, frequent traffic signals, and traffic violations, which can optimize the vehicle driving behavior and improve the safety and efficiency of intersection traffic. By using naturalistic driving data from the DAIR V2X-Seq dataset and general vehicle dynamic parameters, it is possible to obtain the joint-probability-density distribution of the bivariate dynamic parameters of a vehicle. This distribution represents the driving characteristics of vehicles in intersection traffic. The three vehicle dynamic parameters that have an impact on vehicles driving through the intersection area are velocity, angular velocity, and acceleration. The driving behavior characteristics of human-driven vehicles (HVs) and autonomous vehicles (AVs) were analyzed using the multivariate kernel density estimation (MKDE) method to establish the vehicle-behavior boundary. The assessment of the boundary model showed that it accurately characterizes the driving characteristics of HVs and AVs. This boundary can be used to improve the safety detection of intersection areas, enhancing the performance of autonomous vehicles and optimizing intersection traffic.

show abstract

Section: Related Workmentioning

confidence: 99%

Research on the Vehicle-Behavior Boundary of Intersection Traffic Based on Naturalistic Driving Data Study

Wu,

Ma,

Zhu

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

show abstract

Inter- and Intra-Driver Reaction Time Heterogeneity in Car-Following Situations

Tawfeek

2024

Sustainability

View full text Add to dashboard Cite

This study aims to examine the differences in drivers’ reaction time (RTs) while driving on horizontal curves and straight roadway segments, among different driver classes, and in different driving environments to better understand human driver behavior in typical car-following situations. Therefore, behavioral measures were extracted from naturalistic car-following trajectories to estimate the RT. The RT was estimated for two stimulus–response pairs, namely, the speed–gap and relative speed–acceleration pairs, by using the cross-classification method. The RT was estimated separately for each driver and aggregated based on location and based on driver class. The results reveal that drivers’ RTs on curves are consistently higher than their RTs on straight segments, and this difference is statistically significant. The comparison between normal drivers and aggressive drivers indicates that regardless of the location, aggressive drivers have a significantly longer RT than normal drivers, as aggressive drivers can accept closer gaps and higher relative speed. Also, cautious drivers have a longer RT compared with normal drivers; however, the difference is not significant in most cases. Furthermore, cautious and normal drivers have longer RTs on curves compared with their RTs on straight segments. Additionally, the RT on rural horizontal curves is longer than the RT on urban curves, yet the differences are insignificant.

show abstract

Research on The Safety Boundary of Intersection Car-Following Scenarios Based on Naturalistic Driving Behavior Study

Ma¹,

Wei²,

Zhou³

2022

SAE Technical Paper Series

View full text Add to dashboard Cite

Human-like speed modeling for autonomous vehicles during car-following at intersections

Cited by 3 publications

References 46 publications

Research on the Vehicle-Behavior Boundary of Intersection Traffic Based on Naturalistic Driving Data Study

Research on the Vehicle-Behavior Boundary of Intersection Traffic Based on Naturalistic Driving Data Study

Inter- and Intra-Driver Reaction Time Heterogeneity in Car-Following Situations

Research on The Safety Boundary of Intersection Car-Following Scenarios Based on Naturalistic Driving Behavior Study

Contact Info

Product

Resources

About