Traffic microsimulation combined with surrogate safety measures has increasingly been used as a proactive alternative to historical crash data for predicting crash frequency for current or planned road infrastructure designs. However, existing microsimulation-based safety studies have adopted simplified rule-based behaviour models, which reproduce traffic flow reasonably well but often fail to generate realistic conflict dynamics, limiting crash prediction accuracy. Recent advances in machine learning (ML)-based behaviour models offer a promising opportunity to potentially improve microsimulation realism and crash frequency predictions by learning human driving behaviour directly from large-scale trajectory datasets. To investigate this possibility, traffic microsimulation was conducted for five real-world signalised intersections in Leeds, UK, using both a standard rule-based model and a state-of-the-art ML model. Simulated vehicle trajectories were analysed using a two-dimensional Time-to-Collision metric to identify simulated conflicts, which were then modelled using Extreme Value Theory to predict crash frequency. Results show that conflicts from the ML model yielded crash predictions in line with the real-world crash data, whereas the rule-based model did not permit meaningful predictions, presumably due to a lack of model calibration to the specific simulated intersections. Directly using ML-generated simulated crashes to predict real-world crash frequency also yielded poor results, suggesting that while current ML models can realistically reproduce conflicts, they are not yet able to generate realistic crashes. Overall, the findings demonstrate that ML-based behaviour models are promising for improving crash prediction from simulated conflicts, without a need for location-specific model calibration, and suggest clear future directions for ML-based traffic microsimulation.

Safe and efficient interaction with pedestrians is critical for the successful deployment of autonomous vehicles (AVs). While vehicle speed and acceleration are vital communication signals for this interaction, the ways in which human drivers use this kinematic communication signals remain under-explored. To design motion-based communication for AVs, this study aims to explore how drivers control their speed as a function of initial interaction conditions, and how this translates into more or less successful interaction outcomes. We analysed naturalistic trajectory data of vehicles and pedestrians from two zebra crossings in Leeds, UK, comprising 981 interactions. Partial Least Squares Structural Equation Modelling (PLS-SEM) was applied to quantify the mediating effect of vehicle kinematics on the relationship between initial time gap and interaction outcomes (crossing order, efficiency, and safety). Based on these models, vehicle speed profiles associated with optimal and worst outcomes were identified and compared using Generalized Estimating Equations (GEE). PLS-SEM results confirmed that vehicle kinematics significantly mediate interaction outcomes, exerting a strong influence on efficiency and safety, and a weaker influence on crossing order. Minimum vehicle speed in particular exerted a strong influence on crossing order and safety, suggesting that the speed to which the driver drops is a key communication signal to pedestrians. The mediation results and trajectory analysis revealed that drivers used different strategies to achieve balance between their efficiency and interaction safety: in yielding scenarios (time gap > 3 s), they decelerated early and softly to maintain a low non-zero speed (~1 m/s) at a far distance (>10 m) to ensure safety while conserving momentum to ensure efficiency; in ambiguous non-yielding scenarios (time gaps ~ 1 – 3 s), they employed a two-stage strategy: a proactive initial deceleration to prepare for potential danger and reduce uncertainty, followed by an assertive acceleration for efficiency; in unambiguous non-yielding scenarios with a narrow time gap (< 1 s), they sacrificed a small amount of efficiency and adopted proactive deceleration before pedestrians displayed a clear crossing intention to ensure safety. This study expands the understanding of interaction beyond binary yielding choices, offering new insights into the proactive strategies human drivers use to negotiate right-of-way safely and efficiently. Emulating these kinematics strategies can support better integration of AVs into human traffic.

Commuting is a fundamental aspect of daily life; however, its psychological and experiential dimensions remain insufficiently examined within regional contexts, particularly in West Yorkshire, a region undergoing significant transport reforms. This study examined the interrelationships among traffic climate, commuting stress, and travel satisfaction in West Yorkshire, United Kingdom, while accounting for demographic and contextual variations. Utilising survey data from 978 commuters, we evaluated perceptions of traffic climate (Traffic Climate Scale), commuting stress (Multimodal Commuting Stress Scale), and overall travel satisfaction (Scale for Travel Satisfaction). Among different transport modes, car drivers reported higher satisfaction and lower stress compared to public transport users, who experienced the greatest inefficiency and insecurity. Urban residents perceived higher affective demands and psychophysical strain than their rural counterparts. Perceiving the traffic system as functional was positively correlated with travel satisfaction and negatively correlated with commuting stress, whereas external affective demands and internal requirements exhibited the opposite pattern. Hierarchical regression analysis revealed that demographic and contextual factors accounted for only 6% of the variance in travel satisfaction, while traffic climate and commuting stress explained an additional 43%, underscoring their predominant role in shaping commuting experiences. These findings emphasise the importance of enhancing traffic functionality and mitigating emotional and physical stressors to improve commuter travel satisfaction and, as a result, well-being and inform sustainable transport planning.

Autonomous vehicles (AVs) must communicate yielding intent to pedestrians in mixed traffic to ensure safe interaction, and vehicle decelerating motion is a promising implicit communication channel at crosswalks. Designing AV deceleration motion remains difficult because it is continuous with interdependent kinematics (e.g., braking onset, deceleration rate), making fixed‑design experiments inefficient. Therefore, we parameterised deceleration behaviour using a minimum‑jerk model (MJM), which can generate smooth speed profiles from tuneable parameters (stopping distance and duration). We will then formulate profile design as an optimisation problem that maximises pedestrian‑rated safety and efficiency. We will conduct an adaptive experiment in an immersive pedestrian simulator in which Bayesian optimisation sequentially proposes profiles that are most likely to improve safety and efficiency ratings based on pedestrians’ evaluations of previous profiles. Preliminary results from fitting MJM to 62 naturalistic human profiles indicate that the MJM reproduces the overall human profile shape (median RMSE = 0.88 m/s) with only 2 parameters (i.e., stopping distance, duration), supporting its use as a profile generator for the experiment. This work can contribute a human-in-the-loop framework for designing AV motion as an implicit interface.

Autonomous vehicles (AVs) are rapidly advancing and are expected to play a central role in future mobility. Ensuring their safe deployment requires reliable interaction with other road users, not least pedestrians. Direct testing on public roads is costly and unsafe for rare but critical interactions, making simulation a practical alternative. Within simulation-based testing, adversarial scenarios are widely used to probe safety limits, but many prioritise difficulty over realism, producing exaggerated behaviours which may result in AV controllers that are overly conservative. We propose an alternative method, instead using a cognitively inspired pedestrian model featuring both inter-individual and intra-individual variability to generate behaviourally plausible adversarial scenarios. We provide a proof of concept demonstration of this method's potential for AV control optimisation, in closed-loop testing and tuning of an AV controller. Our results show that replacing the rule-based CARLA pedestrian with the human-like model yields more realistic gap acceptance patterns and smoother vehicle decelerations. Unsafe interactions occur only for certain pedestrian individuals and conditions, underscoring the importance of human variability in AV testing. Adversarial scenarios generated by this model can be used to optimise AV control towards safer and more efficient behaviour. Overall, this work illustrates how incorporating human-like road user models into simulation-based adversarial testing can enhance the credibility of AV evaluation and provide a practical basis to behaviourally informed controller optimisation.

INTRODUCTION:There has been a surge in interest in evaluating new forms of communication for Automated Vehicles (AVs), namely external human-machine interfaces (eHMIs). However, much of the research has focused on younger pedestrians' crossing behavior and experience while interacting with AVs and in daytime conditions with optimum visibility. Given that the AVs will interact with pedestrians of all ages, and at all times, there are still key knowledge gaps that need to be addressed. METHOD:Using a cave-based pedestrian lab, this study investigated the effect of AV kinematics (i.e., deceleration, speed, time gaps) and eHMI (a Slow Pulsing Light Band) on the crossing behavior of younger adult pedestrians (18-35 years old) and older pedestrians (64-77 years old), in both daytime and nighttime virtual environments. RESULTS:Results showed that older pedestrians adopted a different crossing strategy than younger pedestrians. If they decided to cross in the non-deceleration trials, they compensated for their longer crossing duration by initiating their crossing earlier than younger pedestrians. However, if they decided to wait until the deceleration was more prominent, they waited longer than the younger pedestrians. Generally, pedestrians reported feeling less safe and behaved more cautiously during nighttime crossings (i.e., less likely to cross, longer crossing initiation time (CIT) when there was no eHMI). eHMI decreased CIT for both age groups but was interpreted and used differently between younger and older pedestrians. Finally, an eHMI failure trial mainly affected younger pedestrians. CONCLUSIONS AND PRACTICAL APPLICATIONS:This knowledge should inform the design of effective communication for AVs for younger and older pedestrians.

Human behaviour is heterogeneous, a feature that is particularly important in Human Factors (HF) research where performance variability can have safety-critical implications. However, HF studies often focus on average effects between experimental conditions, treating individual differences as statistical noise rather than as meaningful information. Modelling behavioural heterogeneity can improve understanding of how systems affect users and support stronger theoretical development. Multilevel Models (MLMs) provide a flexible statistical framework for analysing hierarchical data structures, such as repeated measurements nested within individuals. Advances in statistical software have increased the accessibility of MLMs, yet many HF studies do not fully exploit their ability to model individual differences in experimental effects. One barrier is limited availability of practical tutorials demonstrating how MLMs can be applied to typical HF datasets. This manuscript addresses this gap by providing a practical introduction to MLMs for HF researchers. First, we review MLM principles and their relevance to HF research. We then present two worked examples. Study 1 demonstrates MLMs as an extension of linear regression for continuous predictors; Study 2 applies MLMs to factorial designs and discusses strategies for managing convergence in complex random-effects structures. Analyses are supported by example datasets and code in R, Python, and MATLAB.

Evidence accumulation models provide a formal framework for studying decision making as a dynamic process unfolding over time. While these models have been extensively developed and reviewed in laboratory paradigms, their structured application in complex, ecologically valid domains has received comparatively little attention. Road traffic is a particularly relevant context for studying sustained, embodied perception action behavior, where decisions unfold under time pressure and involve continuous control and ongoing perception-action coupling. Examining how EAMs have been applied in this domain may therefore offer insights beyond discrete laboratory tasks toward decision making in real-world behavior. This semi-systematic review synthesizes 28 studies (2014-2026) applying EAMs to traffic-related behavior. We organize the literature along two dimensions: 1) modelling level, distinguishing models at the level of discrete decision-making and models at the level of continuous action control, and 2) model architecture, distinguishing evidence accumulation as either a stand-alone decision model or an embedded component within broader perception-action or interaction frameworks. These distinctions are associated with systematic differences in model architecture, parameterization, data usage, and validation strategies, reflecting task specific demands. By providing a structured overview of these patterns, this review clarifies how EAMs are currently instantiated in traffic contexts and highlights methodological challenges and future directions both in traffic modelling and in modelling of decision-making more broadly. Promising directions include laboratory work on evidence accumulation in sustained and time-varying tasks, interactive multi-individual decision-making, and the use of neurophysiological measures to identify the perceptual evidence underlying complex perception-action behavior.

As autonomous driving technology advances, automated vehicles (AVs) will increasingly share road space with pedestrians, creating significant challenges for AV systems. Effective interaction between AVs and pedestrians is one of the key hurdles. Pedestrian simulation tools offer the potential to expedite the evaluation and refinement of these interactive capabilities. However, existing research lacks efforts to model pedestrian behavior in vehicle-yielding scenarios, resulting in distorted modeling results. This paper proposes a perceptually plausible road-crossing decision model that creates temporal-dynamic crossing decisions across a range of vehicle-yielding scenarios. Specifically, a proposed hybrid perception strategy explains how pedestrians may apply psychophysical cues to make crossing decisions. Discrete choice models based on the hybrid perception strategy combined with a crossing initiation model reproduce the details of crossing decisions: the decision and its timing. An empirical dataset collected in a pedestrian simulator is applied to validate the model. Additionally, the latest crossing decision models, i.e., the evidence accumulation model and the artificial neural networks approach, are employed as comparisons. The results show that the proposed model accurately reproduces crossing decision patterns affected by diverse vehicle kinematics in vehicle-yielding scenarios in a perceptually plausible manner. Our results strengthen the notion that there is a perceptual threshold for pedestrians to control their decision-making strategy. The proposed theory and approach bring insights into the computational pedestrian road-crossing behavior and have practical implications in traffic simulation and AV development.

In this paper we introduce a general estimation methodology for learning a model of human perception and control in a sensorimotor control task based upon a finite set of demonstrations. The model's structure consists of (i) the agent's internal representation of how the environment and associated observations evolve as a result of control actions and (ii) the agent's preferences over observable outcomes. We consider a model's structure specification consistent with active inference, a theory of human perception and behavior from cognitive science. According to active inference, the agent acts upon the world so as to minimize surprise defined as a measure of the extent to which an agent's current sensory observations differ from its preferred sensory observations. We propose a bi-level optimization approach to estimation which relies on a structural assumption on prior distributions that parameterize the statistical accuracy of the human agent's model of the environment. To illustrate the proposed methodology, we present the estimation of a model for car-following behavior based upon a naturalistic dataset. Overall, the results indicate that learning active inference models of human perception and control from data is a promising alternative to closed-box models of driving.

Advanced Driver Assistance Systems (ADAS) and Automated Driving Systems (ADS) are expected to improve comfort, productivity and, most importantly, safety for all road users. To ensure that the systems are safe, rules and regulations describing the systems' approval and validation procedures are in effect in Europe. The UNECE Regulation 157 (R157) is one of those. Annex 3 of R157 describes two driver models, representing the performance of a "competent and careful" driver, which can be used as benchmarks to determine whether, in certain situations, a crash would be preventable by a human driver. However, these models have not been validated against human behavior in real safety-critical events. Therefore, this study uses counterfactual simulation to assess the performance of the two models when applied to 38 safety-critical cut-in near-crashes from the SHRP2 naturalistic driving study. The results show that the two computational models performed rather differently from the human drivers: one model showed a generally delayed braking reaction compared to the human drivers, causing crashes in three of the original near-crashes. The other model demonstrated, in general, brake onsets substantially earlier than the human drivers, possibly being overly sensitive to lateral perturbations. That is, the first model does not seem to behave as the competent and careful driver it is supposed to represent, while the second seems to be overly careful. Overall, our results show that, if models are to be included in regulations, they need to be substantially improved. We argue that achieving this will require better validation across the scenario types that the models are intended to cover (e.g., cut-in conflicts), a process which should include applying the models counterfactually to near-crashes and validating them against several different safety related metrics.

Understanding driver-pedestrian interactions at unsignalized locations has gained additional importance due to recent advancements in vehicle automation. Naturalistic observations can only provide correlational data, of limited value for understanding and modeling the mechanisms underlying road user interaction. Therefore, controlled studies in virtual reality (VR) are an important complement, but conventional methods can only accommodate a single human participant. Recently, there has been some interest in studying interactions in VR, by means of distributed simulation, involving multiple human participants. However, there is a lack of validation of this method. Here, we provide a validation study, focusing on a distributed vehicle-pedestrian interaction setup, where pairs of one driver and one pedestrian interacted under various kinematic conditions in a connected virtual environment. To test the validity of the distributed simulation, we used a naturalistic dataset collected in the same UK city, at similar locations, and compared the observed behavior between the two settings. Our results indicate a good relative validity of the simulator study, where road users showed similar non-verbal communication behavior in both datasets. As an additional means of validation, we also leveraged a set of game theoretic models that were developed based on the simulator studies, and found that when applied to the naturalistic dataset, we obtained similar (although not identical) model selection results. The findings suggest that distributed simulation can also be useful for development of computational models of interaction. Overall, the findings suggest that distributed simulation can be a highly valuable tool for studying and modeling road user interactions.

This paper presents a model of pedestrian crossing decisions, based on the theory of computational rationality. It is assumed that crossing decisions are boundedly optimal, with bounds on optimality arising from human cognitive limitations. While previous models of pedestrian behaviour have been either 'black-box' machine learning models or mechanistic models with explicit assumptions about cognitive factors, we combine both approaches. Specifically, we model mechanistically noisy human visual perception and assumed rewards in crossing, but we use reinforcement learning to learn bounded optimal behaviour policy. The model reproduces a larger number of known empirical phenomena than previous models, in particular: (1) the effect of the time to arrival of an approaching vehicle on whether the pedestrian accepts the gap, the effect of the vehicle's speed on both (2) gap acceptance and (3) pedestrian timing of crossing in front of yielding vehicles, and (4) the effect on this crossing timing of the stopping distance of the yielding vehicle. Notably, our findings suggest that behaviours previously framed as 'biases' in decision-making, such as speed-dependent gap acceptance, might instead be a product of rational adaptation to the constraints of visual perception. Our approach also permits fitting the parameters of cognitive constraints and rewards per individual, to better account for individual differences. To conclude, by leveraging both RL and mechanistic modelling, our model offers novel insights about pedestrian behaviour, and may provide a useful foundation for more accurate and scalable pedestrian models.

Understanding collision avoidance behavior is of key importance in traffic safety research and for designing and evaluating advanced driver assistance systems and autonomous vehicles. While existing experimental work has primarily focused on response timing in traffic conflicts, the goal of the present study was to gain a better understanding of human evasive maneuver decisions and execution in collision avoidance scenarios. To this end, we designed a driving simulator study where participants were exposed to one of three surprising opposite direction lateral incursion (ODLI) scenario variants. The results demonstrated that both the participants' collision avoidance behavior patterns and the collision outcome was strongly determined by the scenario kinematics and, more specifically, by the uncertainty associated with the oncoming vehicle's future trajectory. We discuss pitfalls related to hindsight bias when judging the quality of evasive maneuvers in uncertain situations and suggest that the availability of escape paths in collision avoidance scenarios can be usefully understood based on the notion of affordances, and further demonstrate how such affordances can be operationalized in terms of reachable sets. We conclude by discussing how these results can be used to inform computational models of collision avoidance behavior.

There has been extensive research on how automated vehicles (AVs) should interact with pedestrians, most of which focused on explicit communication via external Human–Machine Interface (eHMI). However, evidence from human–human interactions shows that pedestrians rarely rely on explicit signals; instead, they predominantly interpret implicit cues from vehicle movements. Prior AV research has generally not designed motion cues to resemble human driving behaviours, which may have exaggerated dependence on eHMIs. This study asks: how would human-like AV kinematics modulate the influence of eHMIs on pedestrian behaviour? In a within-participant design, forty participants completed a cave-based pedestrian simulator experiment examining three deceleration profiles: (1) HH - human-like deceleration pattern and stopping distance; (2) AH - non-human-like deceleration pattern and human-like stopping distance (AH); (3) AA - non-human-like deceleration and stopping distance. The presence of eHMI was also manipulated. Dependent variables include Crossing Initiation Time (CIT), Perceived Safety (PS), Reaction Time to perceiving deceleration onset (RTdec) and Reaction Time to perceiving the eHMI (RTeHMI). Using Bayesian Multilevel Distributional Models, results showed that human-like behaviour (both HH and AH) led to shorter CIT, higher PS ratings compared to AA. HH and AH were largely similar, with only a subtle difference in RTdec. Due to the high saliency of eHMI, HH and AH showed slight additional improvements in CIT and PS compared to without. This study demonstrates that adopting human-like behaviours can substantially enhance AV-pedestrian interaction, without introducing the safety concerns often associated with eHMI, presenting a promising approach for AV communication.

According to crash data reports, most collisions between cyclists and motorized vehicles occur at unsignalized intersections (where no traffic lights regulate vehicle priority). In the era of automated driving, it is imperative for automated vehicles to ensure the safety of cyclists, especially at these intersections. In other words, to safely interact with cyclists, automated vehicles need models that can describe how cyclists cross and yield at intersections. So far, only a few studies have modeled the interaction between cyclists and motorized vehicles at intersections, and none of them have explored the variations in interaction outcomes based on the type of drivers involved. In this study, we compare non-professional drivers (represented by passenger car drivers) and professional drivers (truck and taxi drivers). We also introduce a novel application of game theory by comparing logit and game theoretic models’ analyses of the interactions between cyclists and motorized vehicles, leveraging naturalistic data. Interaction events were extracted from a trajectory dataset, and cyclists’ non-kinematic cues were extracted from videos and incorporated into the interaction events’ data. The modeling outputs showed that professional drivers are less likely to yield to cyclists than non-professional drivers. Furthermore, the behavioral game theoretic models outperformed the logit models in predicting cyclists’ crossing decisions.

Understanding pedestrian behavior is crucial for the safe deployment of Autonomous Vehicles (AVs) in urban environments. Traditional pedestrian behavior models often fall into two categories: mechanistic models, which do not generalize well to complex environments, and machine-learned models, which generally overlook sensory-motor constraints influencing human behavior and which are thus prone to fail in unseen scenarios. We hypothesize that sensory-motor constraints, fundamental to how humans perceive and interact with their surroundings, are essential for realistic simulations. Thus, we introduce a constrained reinforcement learning (RL) model that simulates the crossing decision and locomotion of pedestrians. Our model includes human sensory constraints, giving the agent imperfect information about the environment, and human motor constraints incorporated through a bio-mechanical model of walking. We gathered data from a human-in-the-loop experiment to understand pedestrian behavior. The findings reveal several behavioral patterns not addressed by existing pedestrian models, regarding how pedestrians adapt their walking speed to the kinematics and behavior of the approaching vehicle. Our model successfully captures these human-like walking speed patterns, enabling us to understand these patterns as a trade-off between time pressure and walking effort. Importantly, the model with both sensory and motor constraints performed better than models only incorporating one of the two. Additionally, behavioral patterns related to external human-machine interfaces and light conditions were also captured by the model. Overall, our results not only demonstrate the potential of constrained RL in modeling pedestrian behaviors but also highlight the importance of sensory-motor mechanisms in modeling pedestrian-vehicle interactions.

Modelling pedestrian-driver interactions is critical for understanding human road user behaviour and developing safe autonomous vehicle systems. Existing approaches often rely on rule-based logic, game-theoretic models, or 'black-box' machine learning methods. However, these models typically lack flexibility or overlook the underlying mechanisms, such as sensory and motor constraints, which shape how pedestrians and drivers perceive and act in interactive scenarios. In this study, we propose a multi-agent reinforcement learning (RL) framework that integrates both visual and motor constraints of pedestrian and driver agents. Using a real-world dataset from an unsignalised pedestrian crossing, we evaluate four model variants, one without constraints, two with either motor or visual constraints, and one with both, across behavioural metrics of interaction realism. Results show that the combined model with both visual and motor constraints performs best. Motor constraints lead to smoother movements that resemble human speed adjustments during crossing interactions. The addition of visual constraints introduces perceptual uncertainty and field-of-view limitations, leading the agents to exhibit more cautious and variable behaviour, such as less abrupt deceleration. In this data-limited setting, our model outperforms a supervised behavioural cloning model, demonstrating that our approach can be effective without large training datasets. Finally, our framework accounts for individual differences by modelling parameters controlling the human constraints as population-level distributions, a perspective that has not been explored in previous work on pedestrian-vehicle interaction modelling. Overall, our work demonstrates that multi-agent RL with human constraints is a promising modelling approach for simulating realistic road user interactions.