UX Analysis

Dialogue at the Edge of Fatigue: Personalized Voice Assistant Strategies in Intelligent Driving Systems

With the rapid development of intelligent transportation systems, voice assistants are increasingly integrated into driving environments, providing an effective means to mitigate the risks of fatigued driving. This study explored drivers’ interaction preferences with voice assistants under different fatigue states and proposed a fatigue-state-based dialogue-awakening mechanism. Using Grounded Theory and the Stimulus–Organism–Response (SOR) framework, in-depth interviews were conducted with 25 drivers from diverse occupational backgrounds. To validate the qualitative findings, a driving simulation experiment was carried out to examine the effects of different voice interaction styles on driver fatigue arousal across various fatigue levels. Results indicated that heavily fatigued drivers preferred highly stimulating and interactive voice communication; mildly fatigued drivers tended toward gentle and socially supportive dialogue; while drivers in a non-fatigued state preferred minimal voice interference, activating voice assistance only when necessary. Significant occupational differences were also observed: long-haul truck drivers emphasized practicality and safety in voice assistants, taxi drivers favored voice interactions combining navigation and social content, and private car owners preferred personalized and emotional support. This study enriches the theoretical understanding of fatigue-sensitive voice interactions and provides practical guidance for the adaptive design of intelligent voice assistants, promoting their application in driving safety.

1 version available:

Display Brightness Perception Model Based on Luminance and Chromaticity Spatial Distribution

Year: 2025

Authors: N He, Y Zhang, C Hu, Z Shen, L He, Y Zhang, Joint International Research Laboratory of Information Display and Visualization, School of Electronic Science and Engineering, Southeast University, Nanjing, Jiangsu 210096, China b Shi Cheng Laboratory for Information Display and Visualization, Nanjing, Jiangsu 210008, China c School of Electronic Information Engineering/School of Integrated Circuits, Nanjing Vocational University of Industry Technology, Nanjing, Jiangsu 210023

Quantitative assessment of perceived brightness has been challenging, primarily due to the discrepancy between identical physical luminance in different scenarios and the perceived brightness. This discrepancy fails to accurately convey the actual sensations of brightness as experienced by the human eye. Through a theoretical study of retinal imaging illuminance and its relationship with visual perception characteristics, we determined that physical luminance and pupil size are the primary factors influencing perceived brightness. Furthermore, when the luminance and chromaticity distribution of an image is uneven, and the viewing distance varies, we strategically utilized the spatially weighted corneal flux density from our previous research as an independent variable to characterize perceived brightness. Because it can predict changes in pupil size and incorporates the luminance distribution, it aligns closely with theoretical research. We designed various types of images and conducted perception experiments in dark environments to develop a model for the perception of display brightness. We further selected typical natural images to validate the model's accuracy. The model demonstrates a strong fit and predictive capability of > 0.8, allowing it to predict the perceived brightness of the human eye based on the spatial distribution of luminance and chromaticity of the display. Utilizing this model, we can establish an effective strategy for adjusting display brightness to mitigate the impact of screen light on human eye health.

1 version available:

Driver-Centric Design: Empirical Strategies for Optimizing the Visualization of Driving Information on Vehicle Liquid Crystal Display Dashboards

With the continuous advancement of smart car and vehicle display technologies, liquid crystal display (LCD) dashboards have become the mainstream medium for displaying driving information. This makes the reading performance and user research of LCD vehicle dashboards critically important. In this study, we aim to evaluate the impact of the human-machine interface (HMI) design of vehicle LCD dashboards on driver readability and user experience. Twelve experts participated in a clustering experiment of vehicle dashboards, and 32 drivers participated in a simulated driving environment test of resident display information and temporary display information. The results for resident display information indicate that design type makes a significant difference in both reading performance and visual search efficiency. In addition, only the unconventional-shaped (L-shaped) dual-dial design achieves the ‘‘desired’’ rating in the user experience evaluations. Regarding temporary display information, findings indicate that icons positioned at the top of the dashboard interface are more readable than those at the bottom. Highprioritized or frequently used icons should be placed at the top of the screen. Alternatively, if icons must be positioned at the bottom, they should be no smaller than 8 mm in size. This research will help reduce the driver’s visual workload and off-road sight time, reduce the risk of traffic accidents, and improve the driver’s user experience.

2 versions available

Driving Simulation Performance and Fixational Eye Movement Under Different Photopic and Mesopic Luminance Intensities

The man-machine system is highly involved in driving simulation, as the simulation delivers visual and auditory feedback while drivers' actions, such as braking, become the input. Driving also requires proper visual attention, even under dimmed light intensity. This study compared the driving simulation performance and fixational eye movement under different luminance intensities of tinted lenses and with no tinted lenses. Thirty young adults between the ages of 20 to 28 years participated in this study. All subjects had good vision, proper driving experience, and a license. They were required to complete driving simulation tasks while wearing a headmounted eye tracker under a photopic baseline with no tinted lens, and different tinted lenses of 25%, 50%, and 75% luminance intensity, respectively. Driving simulation performance was evaluated based on course completion duration and braking time. Dikablis eye tracker was used to measure fixational eye movement, which consisted of duration and the number of fixations. The driving simulation performance through course completion duration and braking time differed significantly between different tinted lenses and baseline no lens [F(2.43,70.49)=15.03,p<0.001 and F(2.29, 66.32) = 3.98, p=0.02, respectively]. However, there was no significant difference in the duration and number of fixations between different tinted lenses and baseline no lens [F(2.77, 80.45)=0.56,p=0.63]. This study implied that low light transmission of luminance intensity, especially under mesopic light transmission, affected the drivers' actions in their driving performance, but not their fixational eye movement ability.

1 version available:

Driving Simulator Evaluation of Long Persistent Self-Luminous Pavement Markings’ Visual Guidance

To address the problem of poor nighttime visibility of conventional pavement markings, this study developed a long-persistence self-luminous pavement marking (SLPM) using rare-earth strontium aluminate as the luminescent material. A driving simulator-based virtual environment was employed to evaluate the visual guidance performance of standard and self-luminous markings on both straight and curved highway segments. The analysis focused on gaze behavior indicators, including visual scanning time, scanning angle, and fixation stability. Results show that the long-persistence SLPM markedly enhances drivers’ visual adaptability under low-visibility conditions. Drivers’ visual scanning times were primarily concentrated within the 0–30ms range, with scanning angles between 2° and 4°, while the frequency of scanning was notably lower than that for standard markings, indicating more stable and efficient information acquisition.These findings demonstrate that self-luminous markings can effectively improve nighttime visual guidance, providing theoretical and practical support for the broader implementation of long-persistence luminescent technologies in roadway design.

1 version available:

Dynamic glare evaluation modeling of human eye visual properties and smart materials

In response to the limitations of traditional threshold increment methods in dynamic glare evaluation, this study integrates human visual characteristics with the adaptive properties of intelligent materials to investigate the impact of motion speed on dynamic vision. A fuzzy circle model is employed to simulate the human eye’s refractive effect, analyzing the response of intelligent materials to variations in equivalent luminous screen brightness under different dynamic vision conditions. Dynamic vision detection and road lighting measurement experiments were conducted to validate the proposed approach. Based on these findings, a dynamic glare evaluation model incorporating the sensing and actuation mechanisms of smart materials was developed, enabling adaptive glare perception optimization in response to environmental changes. Experimental results indicate that the relative error between simulated and actual fuzzy circles is below 1%, while the deviation in the dynamic-to-static brightness ratio is only 0.4% in a stationary state, confirming the model’s accuracy and reliability. Additionally, the model exhibits consistency with traditional static glare evaluation methods. This study provides a new theoretical foundation and practical framework for applying intelligent materials in dynamic visual perception assessment.

1 version available:

Enhancing Gaze Prediction in Multi-Party Conversations via Speaker-Aware Multimodal Adaptation

Modeling gaze patterns in multiparty conversations is crucial to build socially-aware dialogue agents and humanoid robots. However, existing approaches typically rely on visual data or focus on dyadic settings. We propose a novel framework for social attention modeling — predicting gaze directions from linguistic and speaker cues alone, without direct visual input. We introduce SAT5, a speaker-aware adaptation of the T5 language model, pre-trained using multi-task objectives that capture both span corruption and speaker state modeling. Using a new dataset of three-party face-to-face conversations with synchronized speech, gaze, and motion capture data, we demonstrate that SAT5 significantly outperforms both pretrained and RNN-based baselines in predicting gaze targets. Our findings highlight the importance of conversational structure and speaker dynamics in modeling social attention, and offer a strong foundation for gaze-aware multimodal systems

1 version available:

Evaluating In-Car Tasks’ Distraction Effects with Drive-In Lab

Existing measurements of driver distraction in laboratory settings lack construct and ecological validity, and therefore, cannot provide reliable estimates of in-car tasks’ distraction effects. In this paper, we operationalize driver distraction in a novel way with the help of Drive-In Lab, where any passenger car can be connected to a driving simulation. The operationalization is based on drivers’ headway maintenance during in-car tasks as compared to baseline driving, while accommodating situational and driver-specific variables, such as brake response times. Realistic visual looming cues enable evaluation of distraction effects on cognitive processes crucial for safe driving. Validation studies with two 2024 car models indicate that the method can reliably differentiate distraction effects between cars, in-car tasks, and drivers as large, medium, small, or no effect on crash potential. The method supports design of in-car interactions by providing valid means to reveal the worst and best practices in in-car user interface design.

1 version available:

Exploring how physio-psychological states affect drivers’ takeover performance in conditional automated vehicles

Although driving automation is promised to improve driving safety, drivers are still required to take over the control of the vehicles in case of emergency. Estimating drivers’ takeover performance serves as the basis for adaptive driving automation and takeover request (TOR) to ensure driving safety. However, although algorithms have been proposed to estimate drivers’ takeover performance through physiological and eye-tracking measures, the complex interrelationships between these metrics and driver behavior, as well as the interactions among the metrics themselves, are not fully understood. To answer this question, a driving simulation experiment involving 42 participants was conducted. Drivers experienced three types of takeover scenarios requested by TOR while driving a conditionally automated vehicle. Drivers’ physiological, eye-tracking metrics and psychological states, as imposed by several non-driving-related tasks were collected. A structural equation model was used to explore the interactions among physiological metrics (i.e., cardiac activity, respiratory activity, electrodermal activity), eye-tracking metrics, psychological states (i.e., trust in driving automation and perceived workload), and variations in takeover time and takeover quality. The results showed that trust was positively associated with takeover quality, while workload was positively associated with takeover time. Additionally, physiological and eye-tracking metrics were indirectly associated with takeover quality via psychological states. This study reveals the hierarchical relationship among takeover-performance-related variables and provides insights for designing driver monitoring systems aimed at estimating takeover performance in vehicles with driving automation and adaptive driving automation to improve driving safety.

2 versions available

External focus strategy improves visuomotor control of gait in older adults

Few studies have adopted external focus strategies to mitigate the negative effects of conscious movement processing in older adults. We investigated whether a single-session intervention (SSI) using an external focus could improve gait stability and visual search behaviors during adaptive locomotion in older adults. Methods We randomly allocated 112 older adults to either an external focus (EXT, n = 56) or a control group (CON, n = 56). Participants performed an obstacle circumvention walking task along an 8-m walkway for five trials at pre-intervention (T0), post-intervention (T1), and retention (T2). The training phase consisted of 20 walking trials with obstacle circumvention. EXT focused on digits displayed on monitors at their path destinations during walking, while CON walked naturally without any specific instructions. Gait kinematics (i.e., gait variabilities and body sway) and visual search data were collected at T0, T1, and T2. Results Only EXT reduced body sway and variability of spatial and temporal gait parameters, while increasing gait speed when comparing T1 and T2 to T0. EXT also reduced the number of visual fixation and fixation duration percentage on the ground while increasing fixation duration percentage on the destination when comparing T1 and T2 to T0. Conclusions This study is the first to explore SSI with an external focus in older adults, providing evidence of significant improvements in gait stability and visual search behaviors that facilitate feedforward planning. Practicing with an external focus strategy could be recommended as an adjunctive psychomotor approach in clinical settings to enhance visuomotor performance in older adults.

2 versions available

Publication Hub Archive

Dialogue at the Edge of Fatigue: Personalized Voice Assistant Strategies in Intelligent Driving Systems

Display Brightness Perception Model Based on Luminance and Chromaticity Spatial Distribution

Driver-Centric Design: Empirical Strategies for Optimizing the Visualization of Driving Information on Vehicle Liquid Crystal Display Dashboards

Driving Simulation Performance and Fixational Eye Movement Under Different Photopic and Mesopic Luminance Intensities

Driving Simulator Evaluation of Long Persistent Self-Luminous Pavement Markings’ Visual Guidance

Dynamic glare evaluation modeling of human eye visual properties and smart materials

Enhancing Gaze Prediction in Multi-Party Conversations via Speaker-Aware Multimodal Adaptation

Evaluating In-Car Tasks’ Distraction Effects with Drive-In Lab

Exploring how physio-psychological states affect drivers’ takeover performance in conditional automated vehicles

External focus strategy improves visuomotor control of gait in older adults