arXiv subjects

Arash Tavakoli

Publications and source records attributed to Arash Tavakoli.

At least 19 recordsLinked to original sources

arXiv2026

When Context Dominates: Multimodal Signatures of Takeover Readiness Under Varying Hazard and Cognitive Load Conditions

Semi-automated driving systems promise to reduce crashes by assisting with perception and control, yet they simultaneously introduce additional human factors challenges by requiring drivers to monitor automation and rapidly resume control when failures occur. Prolonged passive monitoring can degrade vigilance, delay reactions, and increase takeover risk, but the extent to which distraction, hazard context, and drivers' underlying cognitive and physiological states jointly shape takeover performance remains insufficiently understood. This study investigates these interacting factors using a controlled, within-subjects driving simulator experiment that crosses two hazard types (dynamic pedestrian and static crash events) with three levels of secondary task engagement (no task, conversation, and working memory load). Driver responses were assessed using a multimodal sensing framework that integrates vehicle-dynamics measures, subjective workload ratings, autonomic physiology (electrodermal activity and heart rate variability), and prefrontal cortical activation measured with functional near-infrared spectroscopy. Results show that hazard context is the primary determinant of takeover behavior, with pedestrian events producing longer and more variable maneuvers and crash events yielding faster and more stable responses. Secondary tasks exerted smaller effects on objective vehicle control, while internal-state measures showed more variable task-related patterns. These findings highlight the importance of jointly considering environmental context and human state when evaluating takeover readiness and designing driver monitoring systems. This study lays the groundwork for adaptive, context-aware strategies that support safer human-automation collaboration in semi-automated vehicles.

cs.HC↗

arXiv2026

Steering through Time: Blending Longitudinal Data with Simulation to Rethink Human-Autonomous Vehicle Interaction

As semi-automated vehicles (SAVs) become more common, ensuring effective human-vehicle interaction during control handovers remains a critical safety challenge. Existing studies often rely on single-session simulator experiments or naturalistic driving datasets, which often lack temporal context on drivers' cognitive and physiological states before takeover events. This study introduces a hybrid framework combining longitudinal mobile sensing with high-fidelity driving simulation to examine driver readiness in semi-automated contexts. In a pilot study with 38 participants, we collected 7 days of wearable physiological data and daily surveys on stress, arousal, valence, and sleep quality, followed by an in-lab simulation with scripted takeover events under varying secondary task conditions. Multimodal sensing, including eye tracking, fNIRS, and physiological measures, captured real-time responses. Preliminary analysis shows the framework's feasibility and individual variability in baseline and in-task measures; for example, fixation duration and takeover control time differed by task type, and RMSSD showed high inter-individual stability. This proof-of-concept supports the development of personalized, context-aware driver monitoring by linking temporally layered data with real-time performance.

cs.HC↗

arXiv2026

A Multi-Level Data-driven Framework for Understanding Perceptions Towards Cycling Infrastructure Across Regions Leveraging Social Media Discourse

Cycling plays an important role in sustainable urban mobility, yet how people perceive cycling infrastructure varies widely and remains challenging to assess at large spatial scales. Existing research has mainly relied on surveys or short-form social media data and has often focused on individual cities, leaving limited insight into how cycling discussions unfold across broader geographic contexts. This study proposes a multi-scale framework that examines how cycling infrastructure is discussed and evaluated in online public discourse and explores whether sentiment patterns differ between the United States (U.S.) and selected European countries included in the dataset. The analysis draws on a large collection of discussions on a social media platform, namely Reddit, including more than 30,000 posts and over 500,000 associated comments gathered from cycling-focused and geographically defined communities across multiple U.S. states and selected European countries. Using a combination of sentiment analysis, topic modeling, aspect-based classification, and hierarchical statistical modeling, the study evaluates the emotional tone and thematic structure of these discussions and how they vary spatially. Overall sentiment toward cycling is positive in both regions, with slightly higher values observed in the European sample, although differences remain modest. Sentiment tends to become more critical in comment discussions compared to original posts. Topic and aspect analyses show that sentiment is primarily associated with experience-based themes, with most variation occurring within cities rather than between regions. Together, these findings illustrate how discussion-based online data can complement traditional approaches to understanding public perceptions of cycling infrastructure in sustainable urban contexts.

cs.SI↗

arXiv2026

Can Vision-Language Models Understand Construction Workers? An Exploratory Study

As robotics become increasingly integrated into construction workflows, their ability to interpret and respond to human behavior will be essential for enabling safe and effective collaboration. Vision-Language Models (VLMs) have emerged as a promising tool for visual understanding tasks and offer the potential to recognize human behaviors without extensive domain-specific training. This capability makes them particularly appealing in the construction domain, where labeled data is scarce and monitoring worker actions and emotional states is critical for safety and productivity. In this study, we evaluate the performance of three leading VLMs, GPT-4o, Florence 2, and LLaVa-1.5, in detecting construction worker actions and emotions from static site images. Using a curated dataset of 1,000 images annotated across ten action and ten emotion categories, we assess each model's outputs through standardized inference pipelines and multiple evaluation metrics. GPT-4o consistently achieved the highest scores across both tasks, with an average F1-score of 0.756 and accuracy of 0.799 in action recognition, and an F1-score of 0.712 and accuracy of 0.773 in emotion recognition. Florence 2 performed moderately, with F1-scores of 0.497 for action and 0.414 for emotion, while LLaVa-1.5 showed the lowest overall performance, with F1-scores of 0.466 for action and 0.461 for emotion. Confusion matrix analyses revealed that all models struggled to distinguish semantically close categories, such as collaborating in teams versus communicating with supervisors. While the results indicate that general-purpose VLMs can offer a baseline capability for human behavior recognition in construction environments, further improvements, such as domain adaptation, temporal modeling, or multimodal sensing, may be needed for real-world reliability.

cs.CV↗

arXiv2025

A Multimodal Human-Centered Framework for Assessing Pedestrian Well-Being in the Wild

Pedestrian well-being is a critical yet rarely measured component of sustainable urban mobility and livable city design. Existing approaches to evaluating pedestrian environments often rely on static, infrastructure-based indices or retrospective surveys, which overlook the dynamic, subjective, and psychophysiological dimensions of everyday walking experience. This paper introduces a multimodal, human-centered framework for assessing pedestrian well-being in the wild by integrating three complementary data streams: continuous physiological sensing, geospatial tracking, and momentary self-reports collected using the Experience Sampling Method. The framework conceptualizes pedestrian experience as a triangulation enabling a holistic understanding of how urban environments influence well-being. The utility of our framework is then demonstrated through a naturalistic case study conducted in the Greater Philadelphia region, in which participants wore research-grade wearable sensors and carried GPS-enabled smartphones during their regular daily activities. Physiological indicators of autonomic nervous system activity, including heart rate variability and electrodermal activity, were synchronized with spatial trajectories and in situ self-reports of stress, affect, and perceived infrastructure conditions. Results illustrate substantial inter- and intra-individual variability in both subjective experience and physiological response, as well as context-dependent patterns associated with traffic exposure, pedestrian infrastructure quality, and environmental enclosure. The findings also suggest that commonly used walkability indices may not fully capture experiential dimensions of pedestrian well-being. By enabling real-world, multimodal measurement of pedestrian experience, the proposed framework offers a scalable and transferable approach for advancing human-centered urban analytics.

eess.SY↗

arXiv2025

Learning in complex action spaces without policy gradients

While conventional wisdom holds that policy gradient methods are better suited to complex action spaces than action-value methods, foundational work has shown that the two paradigms are equivalent in small, finite action spaces (O'Donoghue et al., 2017; Schulman et al., 2017a). This raises the question of why their computational applicability and performance diverge as the complexity of the action space increases. We hypothesize that the apparent superiority of policy gradients in such settings stems not from intrinsic qualities of the paradigm but from universal principles that can also be applied to action-value methods, enabling similar functions. We identify three such principles and provide a framework for incorporating them into action-value methods. To support our hypothesis, we instantiate this framework in what we term QMLE, for Q-learning with maximum likelihood estimation. Our results show that QMLE can be applied to complex action spaces at a computational cost comparable to that of policy gradient methods, all without using policy gradients. Furthermore, QMLE exhibits strong performance on the DeepMind Control Suite, even when compared to state-of-the-art methods such as DMPO and D4PG.

cs.LG↗

arXiv2025

Simulation for All: A Step-by-Step Cookbook for Developing Human-Centered Multi-Agent Transportation Simulators

As cities evolve toward more complex and multimodal transportation systems, the need for human-centered multi-agent simulation tools has never been more urgent. Yet most existing platforms remain limited - they often separate different types of road users, rely on scripted or pre-defined behaviors, overlook public transit users as active participants, and are rarely designed with accessibility in mind for non-technical users. To address this gap, this paper presents the specifications of a multi-agent simulation platform designed to support real-time, human-centered, and immersive studies of all road users, accompanied by open-source scripts for replication. Using high-fidelity immersive virtual environments, our platform enables interaction across public transit users, pedestrians, cyclists, automated vehicles, and drivers. The architecture is modular, extensible, and designed for accessibility. The system integrates hardware-specific modules - including an omnidirectional treadmill, a seating arrangement, a smart trainer, and an actuated cockpit. Additionally, the platform collects multimodal physiological, neurological, and behavioral data through embedded sensing devices such as functional near-infrared spectroscopy (fNIRS), eye tracking, and wrist-based biosensors. To show the usability of this system, we present three use cases. Simulation for All aims to lower the barrier to entry for high-fidelity transportation simulation, support experimentation across disciplines, and advance our understanding of multimodal mobility in complex urban environments.

cs.MA↗

arXiv2025

Urban Forms Across Continents: A Data-Driven Comparison of Lausanne and Philadelphia

Understanding urban form is crucial for sustainable urban planning and enhancing quality of life. This study presents a data-driven framework to systematically identify and compare urban typologies across geographically and culturally distinct cities. Using open-source geospatial data from OpenStreetMap, we extracted multidimensional features related to topography, multimodality, green spaces, and points of interest for the cities of Lausanne, Switzerland, and Philadelphia, USA. A grid-based approach was used to divide each city into Basic Spatial Units (BSU), and Gaussian Mixture Models (GMM) were applied to cluster BSUs based on their urban characteristics. The results reveal coherent and interpretable urban typologies within each city, with some cluster types emerging across both cities despite their differences in scale, density, and cultural context. Comparative analysis showed that adapting the grid size to each city's morphology improves the detection of shared typologies. Simplified clustering based solely on network degree centrality further demonstrated that meaningful structural patterns can be captured even with minimal feature sets. Our findings suggest the presence of functionally convergent urban forms across continents and highlight the importance of spatial scale in cross-city comparisons. The framework offers a scalable and transferable approach for urban analysis, providing valuable insights for planners and policymakers aiming to enhance walkability, accessibility, and well-being. Limitations related to data completeness and feature selection are discussed, and directions for future work -- including the integration of additional data sources and human-centered validation -- are proposed.

cs.CY↗

arXiv2025

Shaping Passenger Experience: An Eye-Tracking Study of Public Transportation Built Environment

Designing public transportation cabins that effectively engage passengers and encourage more sustainable mobility options requires a deep understanding of how users from different backgrounds, visually interact with these environments. The following study employs eye-tracking technology to investigate visual attention patterns across six distinct cabin designs, ranging from the current and poorly maintained versions to enhanced, biophilic focused, cyclist-friendly, and productivity-focused configurations. A total of N:304 participants engaged with each cabin design while their eye movements such as Fixation Counts, Time to First Fixation (TFF), First Fixation Duration (FFD), Stationary Gaze Entropy (SGE), and Gaze Transition Entropy (GTE) were recorded. Results revealed that alternative cabin configurations consistently exhibited shorter TFFs and lower entropy measures compared to the baseline current version. Specifically, designs incorporating natural elements and biophilic aspects, streamlined layouts, or functional amenities, facilitated quicker orientation and more structured gaze patterns, indicating enhanced visual engagement and possibly reduced cognitive load. In contrast, the poorly maintained cabin design was associated with higher entropy values, suggesting more scattered and less predictable visual exploration. Demographic factors, particularly ethnicity, significantly influenced FFD in certain designs, with Non-white participants showing reduced fixation durations in the enhanced and poorly maintained environments highlighting the importance of inclusive design considerations. Moreover, transportation-related demographic factors such as frequency of public transport use, trip purpose, and duration of use significantly influenced visual attention metrics in various cabin designs.

cs.HC↗

arXiv2024

Public Transit of the Future: Enhancing Well-Being through Designing Human-centered Public Transportation Spaces

Studies show that psychological effects are among one of the top concerns for public transportation users. While many Americans spend a significant portion of their time in public transportation spaces, the impact of the design and maintenance of these spaces on user well-being has not been fully studied. In this study, we conducted a survey to better understand the effect of implementing different designs on people's well-being and perceptual metrics (N=304). Participants were presented with six images depicting different cabin configurations, including (1) the current version of the cabin space, (2) a low-maintenance version, (3) an aesthetically enhanced version, (4) a bike rack-enabled version, (5) a version with an added workspace, and (6) an improved version with biophilic design. After viewing each image, participants' well-being metrics (e.g., stress, and emotion) and their public transportation perception metrics (e.g., perceptions of safety, and reasonable cost) were evaluated. Our results from linear mixed-effect modeling indicated that adding functional amenities and biophilic design elements led to an overall enhancement in well-being and perceptual metrics. Conversely, low maintenance worsened all measured well-being. This research lays the ground for developing human-centered public transportation spaces that can lead to an increase in public transportation adoption.

cs.HC↗

arXiv2023

Unveiling the Impact of Cognitive Distraction on Cyclists Psycho-behavioral Responses in an Immersive Virtual Environment

The National Highway Traffic Safety Administration reported that the number of bicyclist fatalities has increased by more than 35% since 2010. One of the main reasons associated with cyclists' crashes is the adverse effect of high cognitive load due to distractions. However, very limited studies have evaluated the impact of secondary tasks on cognitive distraction during cycling. This study leverages an Immersive Virtual Environment (IVE) simulation environment to explore the effect of secondary tasks on cyclists' cognitive distraction through evaluating their behavioral and physiological responses. Specifically, by recruiting 75 participants, this study explores the effect of listening to music versus talking on the phone as a standardized secondary tasks on participants' behavior (i.e., speed, lane position, input power, head movement) as well as, physiological responses including participants' heart rate variability and skin conductance metrics. Our results show that (1) listening to high-tempo music can lead to a significantly higher speed, a lower standard deviation of speed, and higher input power. Additionally, the trend is more significant for cyclists who had a strong habit of daily music listening (> 4 hours/day). In the high cognitive workload situation (simulated hands-free phone talking), cyclists had a lower speed with less input power and less head movement variation. Our results indicate that participants' HRV (HF, pnni-50) and EDA features (numbers of SCR peaks) are sensitive to cyclists' cognitive load changes in the IVE simulator.

cs.HC↗

arXiv2022

Occupant Privacy Perception, Awareness, and Preferences in Smart Office Environments

Building management systems tout numerous benefits, such as energy efficiency and occupant comfort but rely on vast amounts of data from various sensors. Advancements in machine learning algorithms make it possible to extract personal information about occupants and their activities beyond the intended design of a non-intrusive sensor. However, occupants are not informed of data collection and possess different privacy preferences and thresholds for privacy loss. While privacy perceptions and preferences are most understood in smart homes, limited studies have evaluated these factors in smart office buildings, where there are more users and different privacy risks. To better understand occupants' perceptions and privacy preferences, we conducted twenty-four semi-structured interviews between April 2022 and May 2022 on occupants of a smart office building. We found that data modality features and personal features contribute to people's privacy preferences. The features of the collected modality define data modality features -- spatial, security, and temporal context. In contrast, personal features consist of one's awareness of data modality features and data inferences, definitions of privacy and security, and the available rewards and utility. Our proposed model of people's privacy preferences in smart office buildings helps design more effective measures to improve people's privacy.

cs.HC↗

arXiv2022

Rethinking infrastructure design: Evaluating pedestrians and VRUs' psychophysiological and behavioral responses to different roadway designs

The integration of human-centric approaches has gained more attention recently due to more automated systems being introduced into our built environments (buildings, roads, vehicles, etc.), which requires a correct understanding of how humans perceive such systems and respond to them. This paper introduces an Immersive Virtual Environment-based method to evaluate the infrastructure design with psycho-physiological and behavioral responses from the vulnerable road users, especially for pedestrians. A case study of pedestrian mid-block crossings with three crossing infrastructure designs (painted crosswalk, crosswalk with flashing beacons, and a smartphone app for connected vehicles) are tested. Results from 51 participants indicate there are differences between the subjective and objective measurement. A higher subjective safety rating is reported for the flashing beacon design, while the psychophysiological and behavioral data indicate that the flashing beacon and smartphone app are similar in terms of crossing behaviors, eye tracking measurements, and heart rate. In addition, the smartphone app scenario appears to have a lower stress level as indicated by eye tracking data, although many participants don't have prior experience with it. Suggestions are made for the implementation of new technologies, which can increase public acceptance of new technologies and pedestrian safety in the future.

cs.HC↗

arXiv2022

How are Drivers' Stress Levels and Emotions Associated with the Driving Context? A Naturalistic Study

Understanding and mitigating drivers' negative emotions, stress levels, and anxiety is of high importance for decreasing accident rates, and enhancing road safety. While detecting drivers' stress and negative emotions can significantly help with this goal, understanding what might be associated with increases in drivers' negative emotions and high stress level, might better help with planning interventions. While studies have provided significant insight into detecting drivers' emotions and stress levels, not many studies focused on the reasons behind changes in stress levels and negative emotions. In this study, by using a naturalistic driving study database, we analyze the changes in the driving scene, including road objects and the dynamical relationship between the ego vehicle and the lead vehicle with respect to changes in drivers' psychophysiological metrics (i.e., heart rate (HR) and facial expressions). Our results indicate that different road objects might be associated with varying levels of increase in drivers' HR as well as different proportions of negative facial emotions detected through computer vision. Larger vehicles on the road, such as trucks and buses, are associated with the highest amount of increase in drivers' HR as well as negative emotions. Additionally, shorter distances and higher standard deviation in the distance to the lead vehicle are associated with a higher number of abrupt increases in drivers' HR, depicting a possible increase in stress level. Our finding indicates more positive emotions, lower facial engagement, and a lower abrupt increase in HR at a higher speed of driving, which often happens in highway environments. This research collectively shows that driving at higher speeds happening in highways by avoiding certain road objects might be a better fit for keeping drivers in a calmer, more positive state.

cs.HC↗

arXiv2022

The Impact of Surrounding Road Objects and Conditions on Drivers Abrupt Heart Rate Changes

Recent studies have pointed out the importance of mitigating drivers stress and negative emotions. These studies show that certain road objects such as big vehicles might be associated with higher stress levels based on drivers subjective stress measures. Additionally, research shows strong correlations between drivers stress levels and increased heart rate (HR). In this paper, based on a naturalistic multimodal driving dataset, we analyze the visual scenes of driving in the vicinity of abrupt increases in drivers HR for the presence of certain stress-inducing road objects. We show that the probability of the presence of such objects increases when becoming closer to the abrupt increase in drivers HR. Additionally, we show that drivers facial engagement changes significantly in the vicinity of abrupt increases in HR. Our results lay the ground for a human-centered driving experience by detecting and mitigating drivers stress levels in the wild.

cs.HC↗

arXiv2022

On the Pitfalls of Heteroscedastic Uncertainty Estimation with Probabilistic Neural Networks

Capturing aleatoric uncertainty is a critical part of many machine learning systems. In deep learning, a common approach to this end is to train a neural network to estimate the parameters of a heteroscedastic Gaussian distribution by maximizing the logarithm of the likelihood function under the observed data. In this work, we examine this approach and identify potential hazards associated with the use of log-likelihood in conjunction with gradient-based optimizers. First, we present a synthetic example illustrating how this approach can lead to very poor but stable parameter estimates. Second, we identify the culprit to be the log-likelihood loss, along with certain conditions that exacerbate the issue. Third, we present an alternative formulation, termed $β$-NLL, in which each data point's contribution to the loss is weighted by the $β$-exponentiated variance estimate. We show that using an appropriate $β$ largely mitigates the issue in our illustrative example. Fourth, we evaluate this approach on a range of domains and tasks and show that it achieves considerable improvements and performs more robustly concerning hyperparameters, both in predictive RMSE and log-likelihood criteria.

cs.LG↗

arXiv2022

Orchestrated Value Mapping for Reinforcement Learning

We present a general convergent class of reinforcement learning algorithms that is founded on two distinct principles: (1) mapping value estimates to a different space using arbitrary functions from a broad class, and (2) linearly decomposing the reward signal into multiple channels. The first principle enables incorporating specific properties into the value estimator that can enhance learning. The second principle, on the other hand, allows for the value function to be represented as a composition of multiple utility functions. This can be leveraged for various purposes, e.g. dealing with highly varying reward scales, incorporating a priori knowledge about the sources of reward, and ensemble learning. Combining the two principles yields a general blueprint for instantiating convergent algorithms by orchestrating diverse mapping functions over multiple reward channels. This blueprint generalizes and subsumes algorithms such as Q-Learning, Log Q-Learning, and Q-Decomposition. In addition, our convergence proof for this general class relaxes certain required assumptions in some of these algorithms. Based on our theory, we discuss several interesting configurations as special cases. Finally, to illustrate the potential of the design space that our theory opens up, we instantiate a particular algorithm and evaluate its performance on the Atari suite.

cs.LG↗

arXiv2022

Driver State Modeling through Latent Variable State Space Framework in the Wild

Analyzing the impact of the environment on drivers' stress level and workload is of high importance for designing human-centered driver-vehicle interaction systems and to ultimately help build a safer driving experience. However, driver's state, including stress level and workload, are psychological constructs that cannot be measured on their own and should be estimated through sensor measurements such as psychophysiological measures. We propose using a latent-variable state-space modeling framework for driver state analysis. By using latent-variable state-space models, we model drivers' workload and stress levels as latent variables estimated through multimodal human sensing data, under the perturbations of the environment in a state-space format and in a holistic manner. Through using a case study of multimodal driving data collected from 11 participants, we first estimate the latent stress level and workload of drivers from their heart rate, gaze measures, and intensity of facial action units. We then show that external contextual elements such as the number of vehicles as a proxy for traffic density and secondary task demands may be associated with changes in driver's stress levels and workload. We also show that different drivers may be impacted differently by the aforementioned perturbations. We found out that drivers' latent states at previous timesteps are highly associated with their current states. Additionally, we discuss the utility of state-space models in analyzing the possible lag between the two constructs of stress level and workload, which might be indicative of information transmission between the different parts of the driver's psychophysiology in the wild.

cs.HC↗