Jingyan Wan

Human centric data fusion in Vehicular Cyber-Physical Systems

Building effective Vehicular Cyber-Physical Systems (VCPS) to improve road safety is a non-trivial challenge, especially when we examine how the driver benefits from the existing and proposed technologies in the presence of Human Factors (HF) related negative factors such as information overload, confusion, and distraction. In this paper, we address a human-centric data fusion problem in VCPS. To the best of our knowledge, this work is the first to apply HF to the data fusion problem, which has both theoretical value and practical implications. In particular, we present a new architecture by defining a distinct High-Level (HL) data fusion layer with HF considerations, that is placed between the safety applications on the VCPS and the human driver. A data fusion algorithm is proposed to fuse multiple messages (based on reaction time, message type, preferred evasive actions, severity of the hazards, etc) and to maximize the total utility of the messages. The algorithm is tested with real human drivers to demonstrate the potential benefit of incorporating such human-centric fusion in existing warning systems.

Effects of lead time of verbal collision warning messages on driving behavior in connected vehicle settings

INTRODUCTION Under the connected vehicle environment, vehicles will be able to exchange traffic information with roadway infrastructure and other vehicles. With such information, collision warning systems (CWSs) will be able to warn drivers with potentially hazardous situations within or out of sight and reduce collision accidents. The lead time of warning messages is a crucial factor in determining the effectiveness of CWSs in the prevention of traffic accidents. Accordingly, it is necessary to understand the effects of lead time on driving behaviors and explore the optimal lead time in various collision scenarios. METHODS The present driving simulator experiment studied the effects of controlled lead time at 16 levels (predetermined time headway from the subject vehicle to the collision location when the warning message broadcasted to a driver) on driving behaviors in various collision scenarios. RESULTS Maximum effectiveness of warning messages was achieved when the controlled lead time was within the range of 5s to 8s. Specifically, the controlled lead time ranging from 4s to 8s led to the optimal safety benefit; and the controlled lead time ranging from 5s to 8s led to more gradual braking and shorter reaction time. Furthermore, a trapezoidal distribution of warning effectiveness was found by building a statistic model using curve estimation considering lead time, lifetime driving experience, and driving speed. CONCLUSIONS The results indicated that the controlled lead time significantly affected driver performance. PRACTICAL APPLICATIONS The findings have implications for the design of collision warning systems.

The Effect of Lead Time of Collision Warning Messages on Driver Performance

Collision warning systems (CWSs) are in development in the intelligent transportation system domain to reduce collision accidents. The lead time of warning messages is a crucial factor in determining system effectiveness in the prevention of accidents. The present experiment studied the effects of controlled lead time at 16 levels (predetermined time headway from the subject vehicle to the collision location when the warning message was issued) and lead vehicle conditions (without vs. with lead vehicle) on driving behaviors in various collision scenarios. The results indicated the controlled lead time and lead vehicle conditions significantly affected driver performance. Maximum effectiveness of warning messages was achieved when the controlled lead time was within the range of 4.5s to 6s. When the warning messages were relatively late, the existence of a lead vehicle brought greater safety benefits and more abrupt deceleration. Potential applications of the results in designing of CWSs are further discussed.

Development and Validation of Warning Message Utility Scale (WMUS)

The evaluation of the warning message effectiveness is an important issue in improving communication safety in the system. The goal of the present research was to develop the scale to evaluate the warning message utility, namely, the effectiveness of warning message in preventing accident in general, and an empirical study was conducted to validate the Warning Message Utility Scale (WMUS) in a controlled laboratory environment. The reliability analysis indicated a good the split-half reliability for the WMUS with a Spearman-Brown Coefficient of .873. The predictive validity of WMUS was verified by the significant correlations between the WMUS scores and behavioral indexes of message utility (including reduced kinetic energy and collision rate). The results of regression indicated that the VWMUS is significant predictor of reduced kinetic energy (r2=.339, p<.001) and collision rate (r2=.363, p<.001), which further proved that the validity of WMUS in measuring effectiveness of the warning messages.

The Effects of Vibration Patterns of Take-Over Request and Non-Driving Tasks on Taking-Over Control of Automated Vehicles

ABSTRACT Automated vehicles offer the possibility of significantly increasing traffic safety, mobility, and driver comfort, and reducing congestion and fuel emissions. Current automation technology, however, remains imperfect, and in certain situations, automation will still require the driver to suspend non-driving tasks and take back control of the automated vehicle in a limited period of time. During automated driving, drivers engaged in non-driving tasks (e.g., reading, taking a nap) may not perceive the visual or auditory take-over request in a timely nor accurate manner. Therefore, it is necessary to explore the potential of tactile warning further. This study investigates the effects of vibration patterns of take-over requests (six vibration patterns with different orders of the vibration location) and various realistic non-driving tasks (six non-driving tasks: reading, typing, watching videos, playing games, taking a nap, and monitoring the driving scenario on the driving simulator) on driver take-over behavior, and driver trust and acceptance of automated vehicles. Across all non-driving tasks, the fastest response time was observed with Vibration Pattern 5 (order of the vibration location: back–back–seat–seat). The shortest response time and largest minimum time-to-collision (TTC) also were observed when drivers took back control of the vehicle after monitoring the driving scenario. No interaction effects between vibration patterns and non-driving tasks were observed. Potential applications of the results of designing take-over requests in automated vehicles are discussed.

A human-in-the-loop wireless warning message notification model and its application in connected vehicle systems

ABSTRACT Objective: Vehicle-to-vehicle (V2V) communication has become one of the most active fields of research recently. The implementation of the wireless connected vehicles has widely extended the transmission range of warning messages to inform drivers of hazards ahead. The present study addressed the human component with mathematical modeling of the human reaction time to warning messages in the connected vehicle systems (CVSs) with different confidence intervals (CIs). Methods: In the present study, human performance in warning responses is modeled by extending an existing mathematical model of human performance with the complexity level of tasks. The modeling of human performance with different levels of uncertainty is integrated to propose the warning message notification model in the CVS settings. The warning message notification models were proposed to model the CVSs parameters including maximum available message notification delay, the maximum available machine processing time, the minimum acceptable message notification range, and the designed message display delay. Results: The optimal designs of CVSs parameters were presented in general and for specific conditions by applying the modeling of human performance with different CIs (i.e., 95% and 99% CI) and the warning message notification model with human in the loop. A software interface with the message notification model implemented was presented to discuss the practical benefits of the current work in the design of CVSs.

Development and validation of a model to predict blood alcohol concentrations: Updating the NHTSA equation

OBJECTS To date, multiple models have been developed to estimate blood or breath alcohol concentration (BAC/BrAC). Several factors have been identified that affect the discrepancy between BACs/BrACs and retrospective estimation (eBAC) with existing equations. To the best of our knowledge, a model to quantify the effects of factors on the discrepancy between BAC/BrAC and eBAC is still nonexistent. The goal of this work was to develop a model to provide a more accurate retrospective estimation of breath alcohol concentration (eBAC). METHOD A laboratory study with alcohol consumption and a driving task was conducted with 30 participants (17 male and 13 female) to explore the factors that may contribute to the discrepancy between BrAC and eBAC obtained with existing models. A new eBAC model was developed to improve the estimation of BrAC by modeling effects of gender, weight, and the delay of BrAC measurement on the discrepancy. The validity of the model was tested and established with the data from the experiment conducted in this study and two published research studies, and compared with existing eBAC models. RESULTS Results of the model validity examination indicated that the developed model had higher R squares and lower root-mean-squared errors (RMSE) in estimating BrAC in three experiments compared with the existing eBAC models, including the NHTSA equation, the Matthew equation, the Lewis equation, the Watson equation, and the Forrest equation. CONCLUSION The developed eBAC model had a better performance of BrAC estimation compared with existing eBAC models. The validation of the model with the data from three empirical studies indicated a high level of generalizability in estimating BrAC.

A Novel 2D-3D Hybrid Approach to Vehicle Trajectory and Speed Estimation

We present in this paper a novel surveillance system using calibrated stereo camera pair. The system adopts a 2D-3D hybrid approach where vehicle detection and tracking are first performed in the 2D space. Then, both appearance and depth cues are incorporated into the tracking module based on a hybrid 2D-3D approach. After change detection is carried out in the original image domain, moving vehicles can be detected and tracked over time. Vehicle positions and speeds can be estimated very accurately by first retrieving the 3D coordinates of pixels associated with the tracked vehicles resulting in a 3D vehicle point cloud and fitting a cuboid to it. The proposed stereo vehicle surveillance system is capable of extracting several important driving parameters, including vehicle trajectories, speeds, and orientations. Experimental results have confirmed the efficiency and robustness of the system and its applicability to road safety applications, including speeding and drunk driving detection.