Data-Driven Scenario Specification for AV-VRU Interactions at Urban Roundabouts (original) (raw)
Related papers
IEEE access, 2024
Optimizing traffic dynamics in an evolving transportation landscape is crucial, particularly in scenarios where autonomous vehicles (AVs) with varying levels of autonomy coexist with human-driven cars. While optimizing Reinforcement Learning (RL) policies for such scenarios is becoming more and more common, little has been said about realistic evaluations of such trained policies. This paper presents an evaluation of the effects of AVs penetration among human drivers in a roundabout scenario, considering both quantitative and qualitative aspects. In particular, we learn a policy to minimize traffic jams (i.e., minimize the time to cross the scenario) and to minimize pollution in a roundabout in Milan, Italy. Through empirical analysis, we demonstrate that the presence of AVs can reduce time and pollution levels. Furthermore, we qualitatively evaluate the learned policy using a cutting-edge cockpit to assess its performance in nearreal-world conditions. To gauge the practicality and acceptability of the policy, we conduct evaluations with human participants using the simulator, focusing on a range of metrics like traffic smoothness and safety perception. In general, our findings show that human-driven vehicles benefit from optimizing AVs dynamics. Also, participants in the study highlight that the scenario with 80% AVs is perceived as safer than the scenario with 20%. The same result is obtained for traffic smoothness perception.
Roundabouts: Traffic Simulations of Connected and Automated Vehicles—A State of the Art
IEEE Transactions on Intelligent Transportation Systems
The paper deals with traffic simulation within roundabouts when both "connected and automated vehicles" (CAVs) and human-driven cars are present. The aim is to present the past, current and future research on CAVs running into roundabouts within the Cooperative, Connected and Automated Mobility (CCAM) framework. Both microscopic traffic simulations and virtual reality simulations by dynamic driving simulators will be considered. The paper is divided into five parts. At first, the literature is analysed using the Systematic Literature Review (SLR) methodology based on Scopus database. Secondly, the influence of CAVs on roundabout-specific design features and configuration is analysed. Gap-acceptance models used to define the capacity of the roundabout, one of its most important key performance indicators, are also presented. Third, the most common simulation software are described and analysed in terms of traffic demand implementation. Then the communication approaches and path management algorithms are studied. An example is proposed on the integration of microscopic traffic simulations and dynamic driving simulators virtual reality simulations. Finally, car following models suitable for roundabout traffic are discussed. There is still a gap between simulations and actual experience. There are reasonable doubts on how modelling and optimizing CAVs' behaviour into roundabouts in view of CCAM. It seems that Cooperative, Connected and Automated Vehicles (CCAVs), more than simply Connected and Automated Vehicles (CAVs), could optimise traffic flow, safety and driving comfort within the roundabout. A very promising technology for traffic simulation within the roundabout seems the one based on dynamic driving simulators.
MDDSVsim: An Integrated Traffic Simulation Platform for Autonomous Vehicle Research
Research and development in the field of intelligent transportation systems (ITS) can be costly in terms of both time and money. A significant initial and ongoing investment is often required in order to obtain a physical platform from which experimentation and results may be gained. Simulation of entities, their dynamics and interactions can provide an appropriate and cost effective method for the development of vehicular applications. When simulating traffic behaviour, it is modelled either at a microscopic level, where the individual characteristics and behaviours of each vehicle are reproduced, or at a macroscopic level where the traffic behaviour is aggregated and represented in terms of density, flow and speed. A difficulty with macroscopic simulation it that it often simplifies certain aspects of a scenario under investigation. Non-realistic vehicle dynamics, simplified communication models and idealistic localisation can all detract from the credibility of evaluations carried out. While microscopic simulation can alleviate these concerns, the computational resources required to simulate a large scale scenario, such as a highway, become prohibitive. This paper demonstrates that the integration of a number of simulation platforms can help alleviate the aforementioned concerns. Based on this premise we present MDDSVsim, the integration of (i) VISSIM-a microscopic simulation program for multi-modal traffic flow modelling, (ii) Microsoft Robotics Developer Studio (MRDS)-a robotics simulation platform, (iii) OPNET-a discrete event simulation engine and finally (iv) The World Model, a framework for building perception systems for robots and intelligent vehicles.
Autonomous Navigation in Interaction-Based Environments—A Case of Non-Signalized Roundabouts
IEEE Transactions on Intelligent Vehicles, 2018
To reduce the number of collision fatalities at crossroads intersections, many countries have started replacing intersections with non-signalized roundabouts, forcing the drivers to be more situationally aware and to adapt their behaviors according to the scenario. A non-signalized roundabout adds to the autonomous vehicle planning challenge, as navigating such interaction-dependent scenarios safely, efficiently, and comfortably has been a challenge even for human drivers. Unlike traffic signal-controlled roundabouts, where the merging order is centrally controlled, driving a non-signalized roundabout requires the individual actor to make the decision to merge based on the movement of other interacting actors. Most traditional autonomous planning approaches use rule-based speed assignment for generating admissible motion trajectories, which work successfully in non-interaction-based driving scenarios. They, however, are less effective in interaction-based scenarios as they lack the necessary ability to adapt the vehicle's motion according to the evolving driving scenario. In this paper, we demonstrate an adaptive tactical behavior planner (ATBP) for an autonomous vehicle that is capable of planning human-like motion behaviors for navigating a nonsignalized roundabout, combining naturalistic behavior planning and tactical decision-making algorithm. The human driving simulator experiment used to learn the behavior planning approach and the ATBP design is described in this paper. Index Terms-Adaptive tactical behaviour planner, adaptive control, human factors, naturalistic driving, trajectory planning. I. INTRODUCTION C URRENTLY autonomous or self-driving vehicles are at the heart of academia and industry research because of their multi-faceted advantages that include improved safety, reduced congestion, lower emissions, greater mobility etc. Significant advancement in digital technology (sensing, processing etc.) has pushed the autonomous technology in ground vehicle Manuscript
A bicycle simulator for experiencing microscopic traffic flow simulation in urban environments
2018 21st International Conference on Intelligent Transportation Systems (ITSC)
Urban environments often imply complex transportation infrastructures with manifold different traffic participant using various modes of transport. These traffic participants interact with each other in different ways, often in specific patterns of communication. One option for understanding these interactions may come from microscopic traffic flow simulations. Simulated traffic on modelled urban transportation infrastructures may deliver insights on general traffic-related problems or show specific locations of high risk of accidents or of low traffic quality. Besides having a general view on microscopic traffic flow simulation results, we propose one option for experiencing these simulations from a first-person perspective visualization as one interacting traffic participant on a non-moving physical bicycle. We introduce a procedure for implementing a bicycle simulator for testing various scenarios in three-dimensional environments. By including individual realtime bicycle movements of test subjects into ongoing traffic simulations, we are able to derive individual behavioral strategies to cope with the modelled transportation infrastructure and with simulated vehicle drivers, bicyclists and pedestrians from the point of view of an urban bicyclist. We aim to introduce a novel technique for (1) analyzing present problems of traffic and built infrastructural elements, and, (2) inspecting planned scenarios with variations in traffic compositions (participants and modes) and built infrastructure (inclusion of new design elements). One first test scenario is implemented for gaining first insights on the usefulness of the presented device.
A hierarchical modelling framework for vehicle-bicycle interactions at roundabouts
This paper introduces a framework to model vehicle-bicycle interactions at unsignalized roundabouts. Based on discrete choice theory, a probabilistic model of two hierarchical levels is proposed to represent the driver yielding decision process. The first level models the probability of conflict whereas the second level quantifies the probability of yielding given that a conflict has occurred. A case study is introduced for evaluation of the proposed methodology using real data observed at a typical Swedish roundabout. The results show that the conflict probability is influenced differently depending on the user, cyclist or driver, arriving to the interaction zones. The yielding probability is negatively correlated with the speed of the vehicle when the driver makes decision. The model estimation results also suggest that the relative position of bicycle has larger impacts than its speed i.e. the closer the bicycle is to the conflict zone the greater the impact is on the driver decision. Finally, the empirical analysis also indicates that vehicle speed less than a threshold value shall lead to a high yielding rate, and therefore safer situation at the roundabout.
2013
With recent and important upgrades to North American intersection design guides, roundabouts are gaining popularity as a method of reducing road conflicts, streamlining flow, and curbing excessive speeding of busy intersections. The current design approach, however, makes use of spot-mean speed measures and design criteria which do not take into account yielding behaviour and acceleration/deceleration which may be affected by regional driving culture and local roundabout design. This research paper introduces the methodology being developed for the detailed analysis of driving behaviour, trajectory interpretation, and conflict measures in modern North American roundabouts from video data extracted by means of computer vision. The analysis explores the methods used to prepare microscopic speed maps, compiled speed profiles, lane-change counts, and gap time measures. It also introduces and discusses the interpretation of trajectories at the scale of roundabout merge sections instead of looking at safety from the point of view of a roundabout as a unified system. The research finds significant variation in distributions of speed across five case study roundabouts in the province of Québec, Canada, which may be explained by regional differences in design and road use. It also reports aggressive gap times and uneven traffic flow as a contributing factor to speed.
2021 IEEE International Intelligent Transportation Systems Conference (ITSC), 2021
In a previous study, we presented VT-Lane, a three-step framework for real-time vehicle detection, tracking, and turn movement classification at urban intersections. In this study, we present a case study incorporating the highly accurate trajectories and movement classification obtained via VT-Lane for the purpose of speed estimation and driver behavior calibration for traffic at urban intersections. First, we use a highly instrumented vehicle to verify the estimated speeds obtained from video inference. The results of the speed validation show that our method can estimate the average travel speed of detected vehicles in real-time with an error of 0.19 m/sec, which is equivalent to 2% of the average observed travel speeds in the intersection of study. Instantaneous speeds (at the resolution of 30 Hz) were found to be estimated with an average error of 0.21 m/sec and 0.86 m/sec respectively for free flowing and congested traffic conditions. We then use the estimated speeds to calibrate the parameters of a driver behavior model for the vehicles in the area of study. The results show that the calibrated model replicates the driving behavior with an average error of 0.45 m/sec, indicating the high potential for using this framework for automated, largescale calibration of car-following models from roadside traffic video data, which can lead to substantial improvements in traffic modeling via microscopic simulation.