The introduction of automated driving systems (ADS) presents significant regulatory and operational challenges to ensure safe and responsible deployment in mixed traffic environments. Despite much academic work and efforts of practitioners, these challenges remain open, requiring a transdisciplinary integration of perspectives. This paper draws on insights from a recent transdisciplinary workshop, highlighting the key issues in ADS deployment, including misalignment between regulations and system capabilities, emerging accident types, and gaps in driver understanding and training. Current regulations struggle to keep pace with the advancing capabilities of ADS, resulting in unclear accountability frameworks and inadequate safety measures. The concept of meaningful human control was used as a basis to identify issues. Workshop participants agreed that meaningful human control has an essential role to play to address the identified issues by ensuring that humans can adequately interact with ADS and that ADS are designed in a manner that ensures safe and responsible deployment with clear fail-safes and redundancy mechanisms. The paper advocates for meaningful human control through continuous driver and vehicle assessment, dynamic safety certifications, and stronger communication between regulators and manufacturers to ensure safe and responsible design, regulation and deployment of automated vehicles. Implementing these actions will strengthen ADS regulation and help navigate the ethical and operational complexities of automated driving systems.@en

Detecting abnormal driving behavior is critical for road traffic safety and the evaluation of drivers’ behavior. With the advancement of machine learning (ML) algorithms and the accumulation of naturalistic driving data, many ML models have been adopted for abnormal driving behavior detection (also referred to in this paper as “anomalies”). Most existing ML-based detectors rely on (fully) supervised ML methods, which require substantial labeled data. However, ground truth labels are not always available in the real world, and labeling large amounts of data is tedious. Thus, there is a need to explore unsupervised or semi-supervised methods to make the anomaly detection process more feasible and efficient. To fill this research gap, this study analyzes large-scale real-world data revealing several abnormal driving behaviors (e.g., sudden acceleration, rapid lane-changing) and develops a hierarchical extreme learning machine (HELM)-based semi-supervised ML method using partly labeled data to accurately detect the identified abnormal driving behaviors. Moreover, previous ML-based approaches predominantly utilized basic vehicle motion features (such as velocity and acceleration) to label and detect abnormal driving behaviors, while this study seeks to introduce event-level safety indicators as input features for ML models to improve the detection performance. Results from extensive experiments demonstrate the effectiveness of the proposed semi-supervised ML model with the introduced safety indicators serving as important features. The proposed semi-supervised ML method outperforms other baseline semi-supervised or unsupervised methods as far as various metrics are concerned: for example, it delivers the best accuracy at 99.58% and the best F1-score at 0.9913. The ablation study further highlights the significance of safety indicators for advancing the detection performance of abnormal driving behaviors.
@en

Traffic interactions between merging and highway vehicles are a major topic of research, yielding many empirical studies and models of driver behaviour. Most of these studies on merging use naturalistic data. Although this provides insight into human gap acceptance and traffic flow effects, it obscures the operational inputs of interacting drivers. Besides that, researchers have no control over the vehicle kinematics (i.e., positions and velocities) at the start of the interactions. Therefore the relationship between initial kinematics and the outcome of the interaction is difficult to investigate. To address these gaps, we conducted an experiment in a coupled driving simulator with a simplified, top-down view, merging scenario with two vehicles. We found that kinematics can explain the outcome (i.e., which driver merges first) and the duration of the merging conflict. Furthermore, our results show that drivers use key decision moments combined with constant acceleration inputs (intermittent piecewise-constant control) during merging. This indicates that they do not continuously optimise their expected utility. Therefore, these results advocate the development of interaction models based on intermittent piecewise-constant control. We hope our work can contribute to this development and to the fundamental knowledge of interactive driver behaviour.

@en

Traffic jams occurring on highways cause increased travel time as well as increased fuel consumption and collisions. So-called phantom traffic jams are traffic jams that do not have a clear cause, such as a merging on-ramp or an accident. Phantom traffic jams make up 50% of all traffic jams and result from instabilities in the traffic flow that are caused by human driving behavior. Automating the longitudinal vehicle motion of only 5% of all cars in the flow can dissipate phantom traffic jams. However, driving automation introduces safety issues when human drivers need to take over the control from the automation. We investigated whether phantom traffic jams can be dissolved using haptic shared control. This keeps humans in the loop and thus bypasses the problem of humans’ limited capacity to take over control, while benefiting from most advantages of automation. In an experiment with 24 participants in a driving simulator, we tested the effect of haptic shared control on the dynamics of traffic flow and compared it with manual control and full automation. We also investigated the effect of two control types on participants’ behavior during simulated silent automation failures. Results show that haptic shared control can help dissipating phantom traffic jams better than fully manual control but worse than full automation. We also found that haptic shared control reduces the occurrence of unsafe situations caused by silent automation failures compared to full automation. Our results suggest that haptic shared control can dissipate phantom traffic jams while preventing safety risks associated with full automation.

@en

Autonomous vehicles rely on accurate trajectory prediction to inform decision-making processes related to navigation and collision avoidance. However, current trajectory prediction models show signs of overfitting, which may lead to unsafe or suboptimal behavior. To address these challenges, this paper presents a comprehensive framework that categorizes and assesses the definitions and strategies used in the literature on evaluating and improving the robustness of trajectory prediction models. This involves a detailed exploration of various approaches, including data slicing methods, perturbation techniques, model architecture changes, and post-training adjustments. In the literature, we see many promising methods for increasing robustness, which are necessary for safe and reliable autonomous driving.

@en

Safe and socially acceptable interactions with human-driven vehicles are a major challenge in automated driving. A good understanding of the underlying principles of such traffic interactions could help address this challenge. Particularly, accurate driver models could be used to inform automated vehicles in interactions. These interactions entail complex dynamic joint behaviors composed of individual driver contributions in terms of high-level decisions, safety margins, and low-level control inputs. Existing driver models typically focus on one of these aspects, limiting our understanding of the underlying principles of traffic interactions. Here, we present a Communication-Enabled Interaction model based on risk perception, that does not assume humans are rational and explicitly accounts for communication between drivers. Our model can explain and reproduce observed human interactions in a simplified merging scenario on all three levels. Thereby improving our understanding of the underlying mechanisms of human traffic interactions and posing a step towards interaction-aware automated driving.

@en

The use of partially automated driving systems raises concerns about potential responsibility issues, posing risk to the system safety, acceptance, and adoption of these technologies. The concept of meaningful human control has emerged in response to the responsibility gap problem, requiring the fulfillment of two conditions, tracking and tracing. While this concept has provided important philosophical and design insights on automated driving systems, there is currently little knowledge on how meaningful human control relates to subjective experiences of actual users of these systems. To address this gap, our study aimed to investigate the alignment between the degree of meaningful human control and drivers' perceptions of safety and trust in a real-world partially automated driving system. We utilized previously collected data from interviews with Tesla "Full Self-Driving"(FSD) Beta users, investigating the alignment between the user perception and how well the system was tracking the users' reasons. We found that tracking of users' reasons for driving tasks (such as safe maneuvers) correlated with perceived safety and trust, albeit with notable exceptions. Surprisingly, failure to track lane changing and braking reasons was not necessarily associated with negative perceptions of safety. However, the failure of the system to track expected maneuvers in dangerous situations always resulted in low trust and perceived lack of safety. Overall, our analyses highlight alignment points but also possible discrepancies between perceived safety and trust on the one hand, and meaningful human control on the other hand. Our results can help the developers of automated driving technology to design systems under meaningful human control and are perceived as safe and trustworthy.

@en

The estimation of probability density functions is a fundamental problem in science and engineering. However, common methods such as kernel density estimation (KDE) have been demonstrated to lack robustness, while more complex methods have not been evaluated in multi-modal estimation problems. In this paper, we present ROME (RObust Multi-modal Estimator), a non-parametric approach for density estimation which addresses the challenge of estimating multi-modal, non-normal, and highly correlated distributions. ROME utilizes clustering to segment a multi-modal set of samples into multiple uni-modal ones and then combines simple KDE estimates obtained for individual clusters in a single multi-modal estimate. We compared our approach to state-of-the-art methods for density estimation as well as ablations of ROME, showing that it not only outperforms established methods but is also more robust to a variety of distributions. Our results demonstrate that ROME can overcome the issues of over-fitting and over-smoothing exhibited by other estimators. @en

Unprecedented possibilities of quadruped robots have driven much research on the technical aspects of these robots. However, the social perception and acceptability of quadruped robots so far remain poorly understood. This work investigates whether the way we design quadruped robots' behaviours can affect people's perception of safety in interactions with these robots. We designed and tested a dominant and submissive personality for the quadruped robot (Boston Dynamics Spot). These were tested in two different walking scenarios (head-on and crossing interactions) in a 2x2 within-subjects study. We collected both behavioural data and subjective reports on participants' perception of the interaction. The results highlight that participants perceived the submissive robot as safer compared to the dominant one. The behavioural dynamics of interactions did not change depending on the robot's appearance. Participants' previous in-person experience with the robot was associated with lower subjective safety ratings but did not correlate with the interaction dynamics. Our findings have implications for the design of quadruped robots and contribute to the body of knowledge on the social perception of non-humanoid robots. We call for a stronger standing of felt experiences in human-robot interaction research.

@en

The provision of robotic assistance during motor training has proven to be effective in enhancing motor learning in some healthy trainee groups as well as patients. Personalizing such robotic assistance can help further improve motor (re)learning outcomes and cater better to the trainee's needs and desires. However, the development of personalized haptic assistance is hindered by the lack of understanding of the link between the trainee's personality and the effects of haptic guidance during human-robot interaction. To address this gap, we ran an experiment with 40 healthy participants who trained with a robotic device to control a virtual pendulum to hit incoming targets, either with or without haptic guidance. We found that certain personal traits affected how users adapt and interact with the guidance during training. In particular, those participants with an 'Achiever gaming style' performed better and applied lower interaction forces to the robotic device than the average participant as the training progressed. Conversely, participants with the 'Free spirit game style' increased the interaction force in the course of training. We also found an interaction between some personal characteristics and haptic guidance. Specifically, participants with a higher 'Transformation of challenge' trait exhibited poorer performance during training while receiving haptic guidance compared to an average participant receiving haptic guidance. Furthermore, individuals with an external Locus of Control tended to increase their interaction force with the device, deviating from the pattern observed in an average participant under the same guidance. These findings suggest that individual characteristics may play a crucial role in the effectiveness of haptic guidance training strategies.

@en

This study presents a general optimal trajectory planning (GOTP) framework for autonomous vehicles (AVs) that can effectively avoid obstacles and guide AVs to complete driving tasks safely and efficiently. Firstly, we employ the fifth-order Bezier curve to generate and smooth the reference path along the road centerline. Cartesian coordinates are then transformed to achieve the curvature continuity of the generated curve. Considering the road constraints and vehicle dynamics, limited polynomial candidate trajectories are generated and smoothed in a curvilinear coordinate system. Furthermore, in selecting the optimal trajectory, we develop a unified and auto-tune objective function based on the principle of least action by employing AVs to simulate drivers’ behavior and summarizing their manipulation characteristics of “seeking benefits and avoiding losses.” Finally, by integrating the idea of receding-horizon optimization, the proposed framework is achieved by considering dynamic multi-performance objectives and selecting trajectories that satisfy feasibility, optimality, and adaptability. Extensive simulations and experiments are performed, and the results demonstrate the framework's feasibility and effectiveness, which avoids both dynamic and static obstacles and applies to various scenarios with multi-source interactive traffic participants. Moreover, we prove that the proposed method can guarantee real-time planning and safety requirements compared to drivers’ manipulation.

@en

Understanding behavior of human drivers in interactions with automated vehicles (AV) can aid the development of future AVs. Existing investigations of such behavior have predominantly focused on situations in which an AV a priori needs to take action because the human has the right of way. However, future AVs might need to proactively manage interactions even if they have the right of way over humans, e.g., a human driver taking a left turn in front of the approaching AV. Yet it remains unclear how AVs could behave in such interactions and how humans would react to them. To address this issue, here we investigated behavior of human drivers (N = 19) when interacting with an oncoming AV during unprotected left turns in a driving simulator experiment. We measured the outcomes (Go or Stay) and timing of participants’ decisions when interacting with an AV which performed subtle longitudinal nudging maneuvers, e.g. briefly decelerating and then accelerating back to its original speed. We found that participants’ behavior was sensitive to deceleration nudges but not acceleration nudges. We compared the obtained data to predictions of several variants of a drift-diffusion model of human decision making. The most parsimonious model that captured the data hypothesized noisy integration of dynamic information on time-to-arrival and distance to a fixed decision boundary, with an initial accumulation bias towards the Go decision. Our model not only accounts for the observed behavior but can also flexibly generate predictions of human responses to arbitrary longitudinal AV maneuvers, and can be used for both informing future studies of human behavior and incorporating insights from such studies into computational frameworks for AV interaction planning.@en

Autonomous vehicles currently suffer from a time-inefficient driving style caused by uncertainty about human behavior, which could be improved by accurate and reliable prediction models enabling more efficient trajectory planning. However, the evaluation of such models is commonly over-simplistic, ignoring the asymmetric importance of prediction errors and the heterogeneity of the datasets used for testing. We examine the potential of recasting interactions between vehicles as gap acceptance scenarios and evaluating models in this structured environment. To that end, we develop a framework aiming to facilitate the evaluation of any model, by any metric, and in any scenario. We then apply this framework to state-of-the-art prediction models, which all show themselves to be unreliable in the most safety-critical situations.

@en

In mixed traffic, one of the challenges for autonomous driving technology is how to safe and socially acceptable interaction with human-driven vehicles (HVs). Understanding human cognitive processes during decision-making in interactions with other road users is crucial for enhancing the smooth execution of driving tasks by autonomous vehicles (AVs). This paper proposes a cognitive model of the driver's cumulative information processing based on drift-diffusion model (DDM). By incorporating the initial decision biases, drift rate, and boundary (depending on the initial speed and gaps between ego vehicle and surrounding users) into the existing DDM, our model captures the integrated interaction between individual drivers and other road users. Classic emergency collision avoidance scenarios were constructed based on a driving simulation platform. Our cognitive model accurately described human decision-making in high-risk scenarios, identified key qualitative and quantitative input variables affecting the driver's cognitive processes, and quantified the safety thresholds of the driver's cumulative information processing. Results can support the personalized modeling of human drivers' cognition and facilitate safe and effective interactions between HVs and AVs.

@en

Predicting the future behavior of human road users is an important aspect for the development of risk-aware autonomous vehicles. While many models have been developed towards this end, effectively capturing and predicting the variability inherent to human behavior still remains an open challenge. This paper proposes TrajFlow - a new approach for probabilistic trajectory prediction based on Normalizing Flows. We reformulate the problem of capturing distributions over trajectories into capturing distributions over abstracted trajectory features using an autoencoder, simplifying the learning task of the Normalizing Flows. TrajFlow outperforms state-of-the-art behavior prediction models in capturing full trajectory distributions in two synthetic benchmarks with known true distributions, and is competitive on the naturalistic datasets ETH/UCY, rounD, and nuScenes. Our results demonstrate the effectiveness of TrajFlow in probabilistic prediction of human behavior.

@en

When a person makes a decision, it is automatically accompanied by a subjective probability judgment of the decision being correct, in other words, a confidence judgment. A better understanding of the mechanisms responsible for these confidence judgments could provide novel insights into human behavior. However, so far confidence judgments have been mostly studied in simplistic laboratory tasks while little is known about confidence in naturalistic dynamic tasks such as driving. In this study, we made a first attempt of connecting fundamental research on confidence with naturalistic driver behavior. We investigated the confidence of drivers in left-turn gap acceptance decisions in a driver simulator experiment (N = 17). We found that confidence in these decisions depends on the size of the gap to the oncoming vehicle. Specifically, confidence increased with the gap size for trials in which the gap was accepted, and decreased with the gap size for rejected gaps. Similarly to more basic tasks, confidence was negatively related to the response times and correlated with action dynamics during decision execution. Finally, we found that confidence judgments can be captured with an extended dynamic drift–diffusion model. In the model, the drift rate of the evidence accumulator as well as the decision boundaries are functions of the gap size. Furthermore, we demonstrated that allowing for post-decision evidence accumulation in the model increases its ability to describe confidence judgments in rejected gap decisions. Overall, our study confirmed that principles known from fundamental confidence research extend to confidence judgments in dynamic decisions during a naturalistic task.

@en

Understanding human behavior in overtaking scenarios is crucial for enhancing road safety in mixed traffic with automated vehicles (AVs). Computational models of behavior play a pivotal role in advancing this understanding, as they can provide insight into human behavior generalizing beyond empirical studies. However, existing studies and models of human overtaking behavior have mostly focused on scenarios with simplistic, constant-speed dynamics of oncoming vehicles, disregarding the potential of AVs to proactively influence the decision-making process of the human drivers via implicit communication. Furthermore, despite numerous studies in other scenarios, so far it remained unknown whether overtaking decisions of human drivers are affected by whether they are interacting with an AV or a human-driven vehicle (HDV). To address these gaps, we conducted a “reverse Wizard-of-Oz” driving simulator experiment with 30 participants who repeatedly interacted with oncoming AVs and HDVs, measuring the drivers' gap acceptance decisions and response times. The oncoming vehicles featured time-varying dynamics designed to influence the overtaking decisions of the participants by briefly decelerating and then recovering to their initial speed. We found no evidence of differences in participants' overtaking behavior when interacting with oncoming AVs compared to HDVs. Furthermore, we did not find any evidence of brief decelerations of the oncoming vehicle affecting the decisions or response times of the participants. Cognitive modeling of the obtained data revealed that a generalized drift-diffusion model with dynamic drift rate and velocity-dependent decision bias best explained the gap acceptance outcomes and response times observed in the experiment. Overall, our findings highlight that cognitive models of the kind considered here can provide a generalizable description of human overtaking decisions and their timing. Such models can thus help further advance the ongoing development of safer interactions between human drivers and AVs during overtaking maneuvers.@en

Detecting abnormal driving behavior is critical for road traffic safety and the evaluation of drivers' behavior. With the advancement of machine learning (ML) algorithms and the accumulation of naturalistic driving data, many ML models have been adopted for abnormal driving behavior detection. Most existing ML-based detectors rely on (fully) supervised ML methods, which require substantial labeled data. However, ground truth labels are not always available in the real world, and labeling large amounts of data is tedious. Thus, there is a need to explore unsupervised or semi-supervised methods to make the anomaly detection process more feasible and efficient. To fill this research gap, this study analyzes large-scale real-world data revealing several abnormal driving behaviors (e.g., sudden acceleration, rapid lane-changing) and develops a Hierarchical Extreme Learning Machines (HELM) based semi-supervised ML method using partly labeled data to accurately detect the identified abnormal driving behaviors. Moreover, previous ML-based approaches predominantly utilize basic vehicle motion features (such as velocity and acceleration) to label and detect abnormal driving behaviors, while this study seeks to introduce Surrogate Safety Measures (SSMs) as the input features for ML models to improve the detection performance. Results from extensive experiments demonstrate the effectiveness of the proposed semi-supervised ML model with the introduced SSMs serving as important features. The proposed semi-supervised ML method outperforms other baseline semi-supervised or unsupervised methods regarding various metrics, e.g., delivering the best accuracy at 99.58% and the best F-1 measure at 0.9913. The ablation study further highlights the significance of SSMs for advancing detection performance.@en

Detecting abnormal driving behavior is critical for road traffic safety and the evaluation of drivers' behavior. With the advancement of machine learning (ML) algorithms and the accumulation of naturalistic driving data, many ML models have been adopted for abnormal driving behavior detection. Most existing ML-based detectors rely on (fully) supervised ML methods, which require substantial labeled data. However, ground truth labels are not always available in the real world, and labeling large amounts of data is tedious. Thus, there is a need to explore unsupervised or semi-supervised methods to make the anomaly detection process more feasible and efficient. To fill this research gap, this study analyzes large-scale real-world data revealing several abnormal driving behaviors (e.g., sudden acceleration, rapid lane-changing) and develops a Hierarchical Extreme Learning Machines (HELM) based semi-supervised ML method using partly labeled data to accurately detect the identified abnormal driving behaviors. Moreover, previous ML-based approaches predominantly utilize basic vehicle motion features (such as velocity and acceleration) to label and detect abnormal driving behaviors, while this study seeks to introduce Surrogate Safety Measures (SSMs) as the input features for ML models to improve the detection performance. Results from extensive experiments demonstrate the effectiveness of the proposed semi-supervised ML model with the introduced SSMs serving as important features. The proposed semi-supervised ML method outperforms other baseline semi-supervised or unsupervised methods regarding various metrics, e.g., delivering the best accuracy at 99.58% and the best F-1 measure at 0.9913. The ablation study further highlights the significance of SSMs for advancing detection performance. @en

Recently, multiple naturalistic traffic datasets of human-driven trajectories have been published (e.g., highD, NGSim, and pNEUMA). These datasets have been used in studies that investigate variability in human driving behavior, for example for scenario-based validation of autonomous vehicle (AV) behavior, modeling driver behavior, or validating driver models. Thus far, these studies focused on the variability on an operational level (e.g., velocity profiles during a lane change), not on a tactical level (i.e., to change lanes or not). Investigating the variability on both levels is necessary to develop driver models and AV s that include multiple tactical behaviors. To expose multi-level variability, the human responses to the same traffic scene could be investigated. However, no method exists to automatically extract similar scenes from datasets. Here, we present a four-step extraction method that uses the Hausdorff distance, a mathematical distance metric for sets. We performed a case study on the highD dataset that showed that the method is practically applicable. The human responses to the selected scenes exposed the variability on both the tactical and operational levels. With this new method, the variability in operational and tactical human behavior can be investigated, without the need for costly and time-consuming driving-simulator experiments.

@en