Zachary R. Doerzaph

Investigation of Driver-Infrastructure and Driver-Vehicle Interfaces for an Intersection Violation Warning System

Research was undertaken to design, develop, and evaluate interfaces for a signalized- and stop-controlled violation warning system. Both infrastructure-based warnings (Driver Infrastructure Interface, DII) and vehicle-based warnings (Driver Vehicle Interface, DVI) were considered. The developed interfaces were tested by placing a driver in an instrumented vehicle on a closed test course with a working signalized intersection. The goal of the effort was to determine which DIIs/DVIs were most effective based upon the time to intersection at which the DII/DVI elicited the correct driver response of braking by the stop bar. While the DIIs that were tested were shown to be largely ineffective for violation warning, results showcase the potential of several DVI modalities, by themselves or in combination, to provide effective warnings to a driver violating a signal- or stop-controlled intersection. Furthermore, results indicate that a DVI warning combined with a vehicles enhanced braking capability (brake precharging and panic brake assist) may enhance the range of acceptable DVIs.

Effects of Haptic Brake Pulse Warnings on Driver Behavior during an Intersection Approach

Intersection crashes account for nearly a quarter of all police reported crashes, and 39% of these result in injury or death. In this experiment, haptic warnings were explored as an alternative to auditory and visual warnings as part of an overall effort to reduce the number of intersection related crashes. The study objective was to determine the haptic brake pulse warning candidate that most often results in the driver successfully stopping for an intersection. Five candidate brake pulse warnings were tested; these varied with respect to length and number of pulses. Significant differences were found between haptic conditions for peak and constant deceleration. Participants receiving the haptic warning were 38 times more likely to stop than those receiving no warning.

Assessment of psychophysiological characteristics using heart rate from naturalistic face video data

Heart rate is a strong indicator of a persons psychophysiological state. For this reason, many applications would benefit from noncontact measurement of heart rate. The paper describes a new procedure for estimating blood volume pulse from a video of a persons face, with an emphasis on real-life scenarios. The approach builds on the algorithm known as Eulerian video magnification, which has shown promise under laboratory conditions, but exhibits problems when attempted in naturalistic situations. In particular, problems arise due to movement by the subject, changing illumination conditions, and low-frame-rate video. This paper describes the procedure that we have developed to address some of these problems, including video rates down to 10 frames per second. The procedure has been tested using videos of indoor subjects, as well as drivers of automobiles in naturalistic situations. The paper also shows analysis and comparison of different stress levels using the extracted heart rate information for a driver on the road.

ECG biometric authentication using a dynamical model

This paper concerns the authentication of individuals through analysis of electrocardiogram (ECG) signals. Because the human heart differs physiologically from one person to the next, ECG signals represent a rich source of information that offers strong potential for authentication or identification. We describe a novel approach to ECG-based biometrics in which a dynamical-systems model is employed, resulting in improved registration of pulses as compared to previous techniques. Parameters at the fiducial points are detected using a sum-of-Gaussians representation, resulting in an 18-component feature vector that can be used for classification. Using a publicly available dataset of ECG signals from 47 participants, a classifier was formulated using quadratic discriminant analysis (QDA). The observed mean authentication accuracies were 90% and 97% using 100 beats and 300 beats, respectively. Although tested with standard ECG signals only, we believe that the approach can be extended to other sensor types, such as fingertip-ECG devices.

Naturalistic Driving Study: Alcohol Sensor Performance

This report analyzes the performance of a passive alcohol sensor included in the head unit of the data acquisition system used in the Strategic Highway Research Program 2 (SHRP 2) Naturalistic Driving Study (NDS). Driver impairment is a critical issue in traffic safety, and the ability to identify alcohol-impaired drivers would be valuable for users of the NDS data. The sensor responds to the presence of alcohol in the cabin air. A positive sensor reading can come from many sources: alcohol from the breath of a driver or other occupant, an open container of an alcoholic beverage, aftershave lotion or perfume, windshield wiper fluid, and even some fast food. On the other hand, open windows may dissipate alcohol from an impaired driver’s breath before it reaches the sensor. Thus, the sensor can produce a positive reading when the driver is sober and can produce a negative reading for an alcohol-positive driver. The objective of this report is to evaluate the sensor performance under several scenarios with known driver alcohol levels and to investigate the feasibility of developing an algorithm to identify potentially alcohol-impaired drivers based on the sensor output.

Extracting information from continuous naturalistic driving data: sample applications

The technology and tools used for naturalistic driving data collection have evolved greatly in recent years. Data collection efforts that required a trunk full of equipment and days of installation can now be achieved with data acquisition systems that are about the size of a deck of cards and can be installed in minutes. This evolution has made possible large-scale driving data collection efforts, such as the upcoming Second Safety Highway Research Program Naturalistic Driving Study (SHRP2 NDS). Data from naturalistic studies allow for an unparalleled breadth and depth of driver behavior analysis that goes beyond the quantification and description of driver distraction into a deeper understanding of how drivers interact with their vehicles. This paper describes key aspects of how such studies are designed and executed, and provides some examples of how common types of data are extracted from these naturalistic driving datasets. Specifically, the use of RADAR and speed data are discussed in detail. In addition, a sample architecture for the storage of and access to these vast quantities of driving data and video is provided. Naturalistic driving data have allowed for a transformation in the understanding of driver behavior and, as datasets are expanded to include diverse populations, they will help researchers and automotive engineers in developing novel ways to mitigate and prevent vehicular crashes and their consequences.

Improving Driver Safety through Naturalistic Data Collection and Analysis Methods

The design of a safe transportation system requires numerous design decisions that should be based on data acquired by rigorous scientific method. Naturalistic data collection and analysis methods are a relatively new addition to the engineers toolbox. The naturalistic method is based on unobtrusively monitoring driver and vehicle performance under normal, everyday, driving conditions; generally for extended collection periods. The method generates a wealth of data that is particularly well-suited for identifying the underlying causes of safety deficiencies. Furthermore, the method also provides robust data for the design and evaluation of safety enhancement systems through field studies. Recently the instrumentation required to do this type of study has become much more cost effective allowing larger numbers of vehicles to be instrumented at a fraction of the cost. This paper will first provide an overview of the naturalistic method including comparisons to other available methods. The focus of the paper then shifts to review the evolution of the data acquisition systems (DAS) and methods that have enabled naturalistic data collection. The goal is to provide readers with an understanding of how technology and unique partnerships has allowed the naturalistic data collection method to mature.

Driver Deceleration and Response Time When Approaching an Intersection: Implications for Intersection Violation Warning

It is estimated that as many as 2.7 million crashes per year occur at intersections or are intersection related. These crashes result in over 8,500 fatalities every year and have prompted substantial research of technologies that provide vehicle-based, infrastructure-based, or infrastructure-cooperative Intersection Violation Warnings (IVWs) to drivers. Such a system would use a pre-specified algorithm to identify drivers that have a high likelihood of violating a traffic control device and subsequently warn the driver to stop. However, prior to developing these algorithms, scientists must understand how drivers respond to traffic signals. The current study characterized this driver response in terms of driver deceleration rates and response time. Drivers approached a signalized intersection at 35 mph (56.3 km/h), while their state (baseline, distracted, and willful) was manipulated and the signal phase changed at various distances. Results indicate that the chosen level of deceleration varied with the state of the driver (e.g. distracted) and the distance from the intersection at which the light changes. Response times, however, did not vary based on these factors. Implications of these results are discussed in terms of their applicability toward adapting the performance of existing Forward Collision Warning algorithms for use in IVW applications.

Biometric authentication using photoplethysmography signals

This paper considers signals from the cardiovascular system for possible use in biometric authentication. The signals of particular interest here derive from photoplethysmography (PPG), which refers to the use of illumination-based sensors that are sensitive to volumetric changes as blood travels through the body. Photoplethysmography sensors have been developed for the fingertip and the ear lobe, and they provide a convenient, noninvasive means of measuring heart rate and heart-rate variability. We demonstrate in this paper that PPG-based signals also have the potential to be used for biometric authentication, even though PPG signals appear to convey much less information than their electromagnetic counterparts, electrocardiograms (ECG). Through a novel decomposition into a sum-of-Gaussians representation, we present experimental results that indicate rank-1 accuracies of 90% and 95% with 2 seconds and 8 seconds of PPG test signal data, respectively. To our knowledge, this paper is the first to demonstrate robust PPG-based authentication for subjects with different emotional states.

Influence of In-Vehicle Adaptive Stop Display on Driving Behavior and Safety

In 2012, 683 000 crashes occurred at stop-sign-controlled intersections, with 2434 of those crashes being fatal and composing 5.3% of all fatal traffic incidents in the United States. Roughly 50% of all fatal crashes at stop-sign-controlled intersections involve crossing over (i.e., running) the traffic control device. With the advent of connected-vehicle technology, it is possible to provide a salient in-vehicle adaptive stop display to a driver. This display could alert a driver when he or she will have to stop at an intersection due to oncoming traffic. The same display could also permit drivers to pass through an intersection without stopping when a conflicting vehicle is not present. The purpose of this paper was to evaluate these potential improvements in safety and mobility through an empirical research study. An in-vehicle adaptive stop display was developed and tested on the Virginia Smart Road. Forty-nine drivers were exposed to multiple intersection scenarios they would experience in the real world while using connected-vehicle technology. The scenarios included variations in adjacent traffic, equipment malfunctions, and total equipment failures. There were no indications of a safety detriment to using the adaptive stop display in terms of compliance (likelihood of driver adhering to the information presented), driver complacency, or driver risk taking. Furthermore, the study indicates that, with a higher measured rate of compliance, an in-vehicle adaptive stop display would have a positive impact on safety.