Understanding Pedestrian Behavior in Complex Traffic Scenes

Rasouli, Amir; Kotseruba, Iuliia; Tsotsos, John K.

doi:10.1109/tiv.2017.2788193

Cited by 185 publications

(83 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Pedestrian looks for some visual cues and the vehicle needs to understand that the pedestrian has noticed its presence before eventually crosing. Our evidence contributes to the debate over the relevance of human eye-contact vs vehicle position signalling, consistent with [18], [19] in that our pedestrians display their own intention to cross to the vehicle by turning their head and looking at it, as seen in the n-grams. A head-turn is easier for the vehicle to see than eye contact from a fixed head position, which suggests this event may have a dual function both to passively observe the vehicle and to actively signal intent to it.…”

Section: Discussionsupporting

confidence: 81%

“…Methods of analysis are often performed via video recording, semi-structured interviews and VR recording. Previous work on pedestrian crossing behavior analysis can be found in [18] [19] a novel dataset composed of 650 video-clips for driver-pedestrian interactions in several locations and different weather conditions. The analysis of their data show that attention plays an important role, as in 90% of the time, pedestrians reveal their intention of crossing by looking at the approaching vehicles.…”

Section: A Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Predicting pedestrian road-crossing assertiveness for autonomous vehicle control

Camara

Giles

Madigan

et al. 2018

2018 21st International Conference on Intelligent Transportation Systems (ITSC)

View full text Add to dashboard Cite

Autonomous vehicles (AVs) must interact with other road users including pedestrians. Unlike passive environments, pedestrians are active agents having their own utilities and decisions, which must be inferred and predicted by AVs in order to control interactions with them and navigation around them. In particular, when a pedestrian wishes to cross the road in front of the vehicle at an unmarked crossing, the pedestrian and AV must compete for the space, which may be considered as a game-theoretic interaction in which one agent must yield to the other. To inform AV controllers in this setting, this study collects and analyses data from real-world human road crossings to determine what features of crossing behaviours are predictive about the level of assertiveness of pedestrians and of the eventual winner of the interactions. It presents the largest and most detailed data set of its kind known to us, and new methods to analyze and predict pedestrian-vehicle interactions based upon it. Pedestrian-vehicle interactions are decomposed into sequences of independent discrete events. We use probabilistic methodslogistic regression and decision tree regression-and sequence analysis to analyze sets and sub-sequences of actions used by both pedestrians and human drivers while crossing at an intersection, to find common patterns of behaviour and to predict the winner of each interaction. We report on the particular features found to be predictive and which can thus be integrated into gametheoretic AV controllers to inform real-time interactions.

show abstract

Section: Discussionsupporting

confidence: 81%

Section: A Related Workmentioning

confidence: 99%

Predicting pedestrian road-crossing assertiveness for autonomous vehicle control

Camara

Giles

Madigan

et al. 2018

2018 21st International Conference on Intelligent Transportation Systems (ITSC)

View full text Add to dashboard Cite

show abstract

“…Therefore, a long transition period including mixed traffic of different levels of automation (i.e., manual, partially and fully automated vehicles) is expected (Litman, 2019). One crucial aspect in terms of road safety is the assurance of clear and intuitive communication between vulnerable road users (VRUs) and vehicles (Rasouli, Kotseruba, & Tsotsos, 2018) with a variety of levels of automation, even if the driver is absent to ensure road safety (Ackermann, Beggiato, Schubert, & Krems, 2019).…”

Section: Introductionmentioning

confidence: 99%

Effects of a light-based communication approach as an external HMI for Automated Vehicles - a Wizard-of-Oz Study

Hensch¹,

Neumann²,

Beggiato³

et al. 2020

ToTS

View full text Add to dashboard Cite

Communication between automated vehicles (AVs) and vulnerable road users (VRUs) is highly relevant in coordinating traffic maneuvers and therefore ensuring road safety. Especially in shared spaces such as parking areas, communication is highly important. As automated driving changes the driver's role, communication between different traffic participants will also change. External human-machine interfaces (eHMIs) may enhance safety and effective communication between VRUs and AVs by providing relevant information to other traffic participants if informal communication is insufficient. Hence, a variety of information (e.g., about AVs' driving mode or future maneuvers) is recommended to be communicated to VRUs. Therefore, we investigated the effects of three different light signals, presented by a light bar placed on the test vehicle's roof, as a form of an eHMI: automation mode, starting mode and crossing mode. Moreover, two different driving conditions (i.e., a manual and a simulated automated driving condition) were implemented to investigate the effects between these conditions. Either the driver was visible in the manual condition or the vehicle appeared driverless as a seat suit covered the driver in the simulated automated condition (Wizard-of-Oz design). A total of N = 173 random pedestrians passing by were interviewed and behavioral data were collected from over 1500 pedestrians. Results indicate that participants felt significantly safer during the interaction with the vehicle when a driver was visible. Although VRUs evaluated the general approach of applying light signals as eHMIs for AVs as useful, they assessed the presented light signals as only partially trustworthy and rather unintuitive. Moreover, many participants were unsure of whether the light signals were directed towards them, thus the directedness of light signals should be considered when implementing a light-based eHMI in AVs. Further, moving light signals attracted more attention (i.e., increased pedestrians' head movements towards the vehicle) than a steady or no light signal did. Interestingly, no difference existed between the investigated steady light signal and the baseline condition (i.e., no light signal) regarding head movements towards the vehicle. The results underline the importance of implementing an appropriate eHMI design in AVs.

show abstract

“…Many challenges remain in the development of automated driving technology. Aside from issues associated with developing suitable infrastructure [4] and regulating autonomous cars, technologies currently used in autonomous vehicles have not achieved the level of robustness to handle various traffic scenarios such as varied weather, lighting conditions, road types or environments [5]. In addition, for vehicles driving in a more complex traffic scene, especially in the urban environment, autonomous vehicles also face the additional challenge of how to achieve effective interaction with other road users.…”

Section: Introductionmentioning

confidence: 99%

“…An accuracy of 72% for head orientation estimation and 85% for motion detection is obtained. Rasouli et al [5,13] used AlexNet to extract features related to pedestrian movement and the surrounding environment, extracted t frames continuously to construct feature matrix, and input this into linear SVM to determine whether pedestrians will cross the street. Ghori et al [14] proposed a real-time learning framework based on the relationship between human posture and intention to realize pedestrian intention recognition.…”

mentioning

confidence: 99%

Research on a Pedestrian Crossing Intention Recognition Model Based on Natural Observation Data

Zhang

Liu

Wang

et al. 2020

Sensors

View full text Add to dashboard Cite

Accurate identification of pedestrian crossing intention is of great significance to the safe and efficient driving of future fully automated vehicles in the city. This paper focuses on pedestrian intention recognition on the basis of pedestrian detection and tracking. A large number of natural crossing sequence data of pedestrians and vehicles are first collected by a laser scanner and HD camera, then 1980 effective crossing samples of pedestrians are selected. Influencing parameter sets of pedestrian crossing intention are then obtained through statistical analysis. Finally, long short-term memory network with attention mechanism (AT-LSTM) model is proposed. Compared with the support vector machine (SVM) model, results show that when the pedestrian crossing intention is recognized 0 s prior to crossing, the recognition accuracy of the AT-LSTM model for pedestrian crossing intention is 96.15%, which is 6.07% higher than that of SVM model; when the pedestrian crossing intention is recognized 0.6 s prior, the recognition accuracy of AT-LSTM model is 90.68%, which is 4.85% higher than that of the SVM model. The determination of pedestrian crossing intention parameter set and the more accurate recognition of pedestrian intention provided in this work provide a foundation for future fully automated driving vehicles.

show abstract

Understanding Pedestrian Behavior in Complex Traffic Scenes

Cited by 185 publications

References 23 publications

Predicting pedestrian road-crossing assertiveness for autonomous vehicle control

Predicting pedestrian road-crossing assertiveness for autonomous vehicle control

Effects of a light-based communication approach as an external HMI for Automated Vehicles - a Wizard-of-Oz Study

Research on a Pedestrian Crossing Intention Recognition Model Based on Natural Observation Data

Contact Info

Product

Resources

About