Conference PaperPDF Available

THE POTENTIAL FOR ADAPTIVE SAFETY THROUGH IN-VEHICLE BIOMEDICAL AND BIOMETRIC MONITORING

Authors:

Abstract and Figures

A 2009 study by the National Highway Traffic Safety Administration identified certain medical conditions as contributing factors in crash causation (Hanna 2009). It was found that about 1.3% of all crashes included in the National Motor Vehicle Crash Causation Survey (NMVCCS) were precipitated by driver reported medical emergencies and 84% of the drivers in crashes precipitated by medical emergencies experienced seizures (epileptic and others), blackouts (non-diabetic), and diabetic reaction prior to the crashes. Drivers who had crashes precipitated by medical emergencies were more likely to sustain severe injury (28% for incapacitating injury and death for crashes with medical emergency; 11% for crashes without medical emergency). Thus, the premise exists that there may be benefit to identify the driver (and other occupants) of the vehicle as well as monitor their current health status through passive or active methods. This monitoring could take into account chronic conditions (such as bone mineral density) through driver input or through initial vehicle startup measurements which could be used to provide optimal comfort or safety system performance. Additional information about the driver's health or behavioral conditions could be interpreted from blood pressure, heart and respiration rate, blood glucose levels and other physiological parameters and could lead to vehicle intervention in driving and/or alert EMS or police of the impending health condition that may affect driving or cause a crash. This monitoring could be done in many ways such as the recent rapid growth in wearable technology with the ability to pair to apps. This paper will discuss issues related to driver behavioral and health monitoring and review potential technologies for monitoring and as well as methods for biometric identification. Recent publications on driver crash risk due to chronic and acute health conditions will be summarized. Finally, applications that may be associated with the monitoring will be discussed.
Content may be subject to copyright.
Ridella 1
THE POTENTIAL FOR ADAPTIVE SAFETY THROUGH IN-VEHICLE BIOMEDICAL AND
BIOMETRIC MONITORING
Stephen A. Ridella
Julie J. Kang
National Highway Traffic Safety Administration
USA
Satoshi Kitazaki
University of Iowa
USA
Paper Number15-0377
ABSTRACT
A 2009 study by the National Highway Traffic Safety Administration identified certain medical conditions as
contributing factors in crash causation (Hanna 2009). It was found that about 1.3% of all crashes included in the
National Motor Vehicle Crash Causation Survey (NMVCCS) were precipitated by driver reported medical
emergencies and 84% of the drivers in crashes precipitated by medical emergencies experienced seizures ( epileptic
and others), blackouts (non-diabetic), and diabetic reaction prior to the crashes. Drivers who had crashes precipitated
by medical emergencies were more likely to sustain severe injury (28% for incapacitating injury and death for
crashes with medical emergency; 11% for crashes without medical emergency). Thus, the premise exists that there
may be benefit to identify the driver (and other occupants) of the vehicle as well as monitor their current health
status through passive or active methods. This monitoring could take into account chronic conditions (such as bone
mineral density) through driver input or through initial vehicle startup measurements which could be used to provide
optimal comfort or safety system performance. Additional information about the driver’s health or behavioral
conditions could be interpreted from blood pressure, heart and respiration rate, blood glucose levels and other
physiological parameters and could lead to vehicle intervention in driving and/or alert EMS or police of the
impending health condition that may affect driving or cause a crash. This monitoring could be done in many ways
such as the recent rapid growth in wearable technology with the ability to pair to apps.
This paper will discuss issues related to driver behavioral and health monitoring and review potential technologies
for monitoring and as well as methods for biometric identification. Recent publications on driver crash risk due to
chronic and acute health conditions will be summarized. Finally, applications that may be associated with the
monitoring will be discussed.
INTRODUCTION
Driver state is an important factor affecting safe driving behavior. Detection and intervention of drowsiness,
distraction and drunkenness have been studied by many up to now. The decline of an individual driver’s health is
another potential cause for a significant proportion of crashes. Researchers as early as 1967 recognized medical
impairment as a possible contributor to traffic “accidents” (Waller, 1967). In his report, Waller outlined seven
criteria that should be adhered to in a study of this subject, including “there must be a reasonable mechanism for
identifying most of the high risk persons, and for doing so early enough to avoid a substantial portion of their
accident experience”. In a recently released report, the National Highway Traffic Safety Administration reviewed
crashes from the National Motor Vehicle Crash Causation Survey (NMVCCS) to determine the “critical reason,
which is the last event in the crash causal chain, and concluded that driver was that reason for 94% of crashes in that
nationally represented survey (NHTSA, 2015). Of those 94% of crashes (representing over 2 million crashes), seven
percent, or 145,000 crashes were attributed to a driver “non-performance error”, which could be drowsiness or an
acute medical condition. Using the same data set, Hanna (2009) found that 1.3% of all crashes involving light
passenger vehicles in NMVCCS were precipitated by a driver’s medical emergency and 84% of the drivers in
crashes precipitated by medical emergencies had experienced seizures (related to epilepsy and other conditions),
blackouts (non-diabetic), and diabetic reaction prior to the crashes (Figure 1).
Hanna found that the drivers who had
crashes precipitated by medical emergencies were more likely to be more severely injured (i.e. 28% suffered
incapacitating injury and death in crashes with a medical emergency compared to 11% in crashes without medical
Ridella 2
emergency). Although the ratio of the crashes
related to acute medical emergencies was small in the study, there
were many other disease-related cognitive and psychomotor impairments (chronic) that may have increased the risk
of crash. The Federal Motor Carrier Safety Administration (FMCSA) commissioned a series of studies to determine
risk of motor vehicle crash for a variety of medical disorders based on reports and medical expert panel opinions
provided by) (Table 1, FMCSA, 2007-2011). The table shows a statistically significant increased risk for a crash
based on a driver having one of several disorders. Obstructive sleep apnea (OSA) is shown to have the highest
relative risk (1.30-5.72). Another general consideration is that some diseases such as type II diabetes, cardiovascular
disease, mild cognitive impairment (MCI) and Alzheimer’s disease are associated with aging. Now that the aging
population is growing rapidly in the US, crash rates of occupants with those diseases may see increases despite the
recent report by the Insurance Institute for Highway Safety (IIHS) indicating that fatal crash involvement rates of
older drivers were decreasing faster than those of younger drivers (IIHS Status Report, 2014). NHTSA’s 5-year plan
for traffic safety for older people (NHTSA, 2013) also highlights older drivers’ risk in association with increased
medical problems.
Advances in vehicle electronics have made it possible for drivers and passengers to customize their driving
experience in many ways. Personalized settings for seat position, as well as heating/cooling/ventilation and
entertainment (separated for drivers and passengers) have been available for many years. Recent technological
progress in diversity, sensitivity, data capacity and miniaturization of biometric and biomedical instrumentation
sensors and devices are enabling the general public to have more real-time access to personal health status as well as
enjoy more security for their personal electronic devices. Recent development has resulted in devices that can be
embedded anywhere such as clothes, wristbands, watches, vehicle interiors, etc. to detect and report medical
information such as body temperature, heart, respiration and perspiration rates, blood glucose and oxygenation
levels, and other physiological functions. This data combined with user identification through recognition of
fingerprints, iris, facial and/or voice inputs can provide a rapid analysis of a person’s state of health. The availability
and use of this information has implications in many markets and significant potential to increase driving comfort
and safety when embedded into appropriate algorithms related to vehicle design and performance. For example, a
vehicle that senses that the driver has an elevated body temperature and has increased his/her respiration rate
significantly may automatically open windows or increase interior ventilation to improve comfort. Also, providing
input to the vehicle that the driver is 75 years old and thus has reduced bone mineral density, the vehicle may adjust
restraint system parameters to optimize occupant protection in the event of a crash. The same vehicle system could
also forward the driver’s vital information to first responders and other health professionals even before they
reached the scene of the crash.
Figure 1 . Type of medical conditions which precipitated crashes (Hanna, 2009).
This paper will present the current state of passive personal identification and monitoring of a person’s health status,
as well as the expected developments of such systems in the future. A discussion of how these devices could
influence vehicle comfort and safety will be provided through a summary of the technology available or in
development, the challenges of integrating the devices to the vehicle, the potential use, accuracy, standardization and
privacy of data as well as other policy implications of this technology. Practical examples will be given to exemplify
Ridella 3
the market readiness of technology and the potential for diversification of products and applications and their
breadth and depth.
HEALTH CONDITIONS THAT MAY AFFECT DRIVING
Medical conditions are generally divided into two categories, chronic and acute problems. Some diseases have either
chronic or acute problems and some have both. Effective in-vehicle interventions for these two aspects are different.
The driver is at extremely high risk of crash in the case of acute health decline which shows distinctive physiological
and behavioral changes. An in-vehicle system that detects a decline and controls the vehicle for an unresponsive
driver may help to to avoid a crash. On the other hand, chronic health problems develop slowly and degrade
cognitive and behavioral performance of the driver over time, resulting in higher general risk of crash. In-vehicle
systems that increase the safety margin of the vehicle based on individual’s driving ability may minimize the
increased risk. The following discussion describes medical conditions that may affect driving.
Table 1.
Summary of motor vehicle crash risk for various medical disorders
(FMCSA, various reports 2007-2011,
Rizzo, 2011)
Medical Condition Relative Risk (95% Confidence Interval)
Cardiovascular Disease 1.43 ( 1.11-1.84)
Diabetes Mellitus 1.28 ( 1.12-1.47)
Obstructive Sleep Apnea 1.30-5.72 (pooled studies)
Seizure Disorders 1.13-2.16 (pooled studies)
Traumatic Brain Injury 1.32 (0.77-2.25)
Diabetes
The likelihood that a person has diabetes (type II) increases with age. Wild et al., (2004) estimated that the number
of people over 65 years old with diabetes in developed countries will increase by 80% between 2000-2030 and the
total number of people over the age of 46 with diabetes will be nearly 80 million people by 2030. An acute risk
factor for driving is hypoglycemia (it can lead to coma) which can be caused mainly by insulin-dependent Type 1
diabetes, whereas the chronic risk factors for driving are neuropathy (decreased sensation at feet and hands),
retinopathy (vision loss), and encephalopathy (cognitive decline), all of which can be caused by both Type 1 and
Type II diabetes (Rizzo,2011). However, scientific evidence for drivers with diabetes being at greater risk for
crashes is not conclusive. Tregear et al. (2007) reviewed and conducted meta-analysis of thirteen studies, comparing
crash risk among drivers with diabetes to drivers without diabetes, and found that the risk for crash among drivers
with diabetes was 19% greater than the risk among drivers without diabetes (within the range published by FMCSA
in Table 1). On the other hand, Tregear et al. found no statistically significant evidence to suggest that insulin-
treated individuals are at higher risk for crash than individuals with diabetes not being treated with insulin. It seems
that diabetes increases the crash risk but contributions from the acute symptoms and the chronic symptoms to the
increased risk are not yet known. The American Diabetes Association (ADA) states that diabetes management and
education of both patients and health care professionals is the important intervention to the driving risk due to
hypoglycemia (American Diabetes Association, 2012).
Obstructive sleep apnea (OSA)
There is strong evidence that the highest relative risk of motor vehicle crash is for OSA, ranging between 1.30 and
5.72 (FMCSA, 2007, Table 1). There is evidence that OSA affects a significant portion of the population regardless
whether it is diagnosed or undiagnosed. Hiestand et al. (2006) conducted a telephone interview of 1506 US adults
using the Berlin questionnaire and found that 26% of respondents met the criteria indicating a high risk of OSA.
NHTSA estimated that 1.4% of total crashes and 1.75% of fatal crashes were related to sleepiness (NHTSA, 1985).
However, Leger (1994) suggested that NHTSA’s estimation was underestimated and provided a new estimation of
sleep-related crash rates as 41.6% of total crashes and 36.1% of fatal crashes. Because of potential under-reporting
and inability to determine post-mortem that drowsiness or micro-sleep episodes contributed to the crash, there is no
clear-cut estimation of crash rate related to OSA. However, considering the high risk factors which were estimated
Ridella 4
in strictly controlled driving simulator studies, OSA should be one of the conditions that can contribute to a risk of a
vehicle crash.
Another important aspect of OSA is the fact that the disease remains undiagnosed in many individuals. Hiestand et
al. (2006) concluded that the prevalence of OSA in the US was estimated to be between 5% and 10%, but only 1 in
10 of those with OSA were adequately diagnosed and screened for Continuous Positive Airway Pressure (CPAP)
treatment. Untreated OSA can cause daytime somnolence, cognitive impairment, loss in work productivity with a
typical symptom of microsleep and increase risk of motor vehicle crashes. It should be also noted that some sleepy
drivers are not aware of their impaired status, possibly because of related cognitive impairment of an altered frame
of reference for fatigue (Rizzo, 2011).
Effects of OSA on driving performance have been investigated in driving simulator experiments by some
researchers. It was found by Paul at al. (2005) and Boyle et al. (2008) that untreated OSA patients showed greater
variation in steering, lane position and TLC (Time-to-Lane-Crossing) during micro-sleep episodes, and degree of
driving performance decrement was correlated with microsleep duration, particularly on curved roads. Risser and
Ware (1999) found that untreated OSA patients demonstrated increased lane position variability and road departure
incident which were positively correlated with frequency and duration of attention lapses (sleeps). Drowsy driver
detection and alert systems are commercially available now. However, it remains unknown to what extent the
current technology detects the critical state of drivers with OSA featured by frequent occurrence of micro-sleep
episodes and to what extent the technology can mitigate the risk of crash.
Other Disease Conditions
In addition to insulin dependent diabetes (typically Type I) and OSA, other diseases such as cardiovascular disease,
seizure disorders and traumatic brain injury can also expose the driver to an increased crash risk due to acute
symptoms (e.g.: heart attack and stroke, epileptic seizure). With cardiovascular disease, there are increasingly more
middle-age and older drivers being treated for symptoms associated with atrial fibrillation and congestive heart
failure. Also, conditions associated with respiratory health such as Chronic Obstructive Pulmonary Disease and
asthma are becoming increasingly common in the adult population. These conditions can turn from chronic to acute
without warning. Each acute health decline is associated with distinctive physiological and behavioral changes from
the normal condition and it is important to have the ability to detect such declines of the occupant while in the
vehicle to activate an in-vehicle intervention.
More long-term degenerative conditions include osteoporosis and mild-to-severe cognitive degeneration such as
Alzheimer’s disease. Ridella et al. (2012) found that osteoporosis or poor bone quality was the most significant
contributing factor to injury, specifically, incidence of rib fractures, in older occupants involved in an injury-
producing car crash. These crashes typically involved lower crash speeds than did crashes involving younger injured
drivers and occupants. While treatable with many medications, bone quality continues to diminish with age with
evidence that the pace of bone loss is more significant in women than men. Cognitive impairments are more difficult
to determine in real-time without an adequate baseline or history for comparison.
REVIEW OF CURRENT TECHNOLOGIES USING DRIVER MONITORING
Occupant Identification
A first step in the process of monitoring is passive, non-invasive identification of the driver and perhaps passengers.
The driver’s identification would be useful in a host of different applications. Establishing identity may allow for
the vehicle to create a baseline of the occupant’s health status that can be used in current as well as future driving
tasks. Algorithms may be developed for the vehicle to learn how it is driven in certain situations and the associated
physiological measurements specific to that person. Also, there may be situations where several drivers share a
single vehicle such as the use of a family car, where there may be a range of driving abilities. A teenager who is
enrolled in a graduated licensing program may have certain vehicle restrictions put upon them whereas other
members of the family may enjoy the full privileges of driving. Therefore, based on the driver’s identity, a
monitoring system would adjust the vehicle’s abilities or monitor highly complex driving tasks more closely.
Ridella 5
In addition, a history of the driver’s performance would also lend insight into the longitudinal data for comparative
purposes. This information could be used to determine long-standing trends in performance such as declining mental
capacities (e.g. Alzheimer’s or dementia), health issues, changes in driving performance, etc.
Technology to determine identity has grown exponentially in the past few years due to consumer demand for
greater security of data. Much of this technology has the ability to be adapted to the vehicle environment or brought
in through portable devices. Camera-based technology can be used for identification through facial recognition or
iris analysis software. Fingerprint, vascular pattern and voice-recognition scanning are other methods of uniquely
identifying an individual. An alternative to passive identification is the driver’s userprovided information through a
key card, implantable radiofrequency ID or RFID or other device that uniquely identifies an individual.
Behavioral Monitoring (DrIIVE)
Driver Monitoring of Impairments and Inattention using Vehicle Equipment (DrIIVE) is a current NHTSA project
that uses driver monitoring data. DrIIVE is focused on the development of an algorithm that can accurately identify
and distinguish among different forms of inattention or impaired driving including alcohol-impaired, drowsy, and
distracted driving. DrIIVE determines driver behavior data from vehicle measures such as steering and pedal inputs,
lane variability, and compares signatures of normal driving with impaired driving. The goal is to use the DrIIVE
algorithm to identify and evaluate the effectiveness of driver monitoring countermeasures on impaired driving
behavior.
Alcohol Impairment (DADDS)
In 2008, NHTSA launched a cooperation program to develop in-vehicle technology that could accurately, precisely,
and reliably measure a driver’s blood alcohol concentration. in a non-invasive way in a very short time. (Monk,
2012). Now in it’s second phase, two subsystems have been developed and are being integrated into a research
vehicle for further testing. One system is breath based and continually samples the area around the driver for alcohol
and carbon dioxide through an infra-red sensor whose measurements can be converted into a blood alcohol
concentration. A second subsystem is touch-based and can measure the absorbed near infra-red light in a person’s
finger and derive an alcohol concentration. These systems have the dramatic potential to reduce crashes and
fatalities involving drunk drivers by denying the driver the ability to start and drive the vehicle.
Physiology/Health Monitoring
Monitoring driver health should be non-invasive and passive. Both in-vehicle and wearable technology have been
developed, however, only wearable technology has been commercialized to date. As sensor technology has become
smaller and less expensive, a vast array of sensor applications have been developed or published. Ford Motor
Company, working with a restraint and sensor supplier, developed a prototype vehicle to measure a variety of
physiological signals (Watson et al, 2011). Figure 2 below indicates that they were looking at both comfort
(temperature difference) and real-time physiology (heart and respiration rate). They indicated that the signals could
be integrated for use in assessing driver performance as a function of wellness, workload, and stress. Ford also has
demonstrated an in-vehicle glucose monitoring system that could detect a driver’s possibility for a diabetic episode
(Ford, 2011).
Demonstrations of blood oxygenation measurements using a variant of typical finger-tip pulse oximeters, Meditech
2011), blood pressure and bone mineral density bring more information about driver/occupant health into the
vehicle. The BOSCOS (BOneSCanning for Occupant Safety, Hardy et al, 2005) project created an in-dashboard
ultrasound sensor that could deduce bone strength based on measurements taken from the distal third of a finger
when inserted into the device.
Wearable devices or mobile human health monitoring is a maturing area of health awareness particularly in the
sports medicine market. Miniaturized electronics or MEMS technology has allowed for creation of wearable wrist
bands, head bands, even undergarments that are capable of accurately measuring heart rate, respiration, sweat
production, etc. These devices are usually coupled to a portable electronic device application to record daily exercise
results and associated physiological responses. The applications have algorithms to detect medical issues and
Ridella 6
performance progress with the person wearing the technology as well as determine stress or anxiety levels. The
cost of these devices has dropped dramatically in the past several years with many devices selling in the $50-$150
range. The sophistication, accuracy and reliability of these devices is steadily improving such that some diagnosis of
heart and respiratory health may be deduced from the signals rather than just the instantaneous rates that are reported
by the devices.
Figure 2. Examples of prototype health monitoring sensors embedded in vehicle components (Wired , 2011)
Establishing a Baseline
Another new challenge within the scope is connection and integration of monitoring the state of the driver in the
home (off board) and in the vehicle (on board). Monitoring a person’s life in the home through so called “Smart
Home” or “Wearable sensors” is becoming important for “Aging in Place” or “Tailor-made Health Care” and it is
actively being tested by research projects for its effectiveness and clarification of system and social requirements.
However, there are no systems / research projects focusing on integration of on and off board driver monitoring.
Early signs of an acute health problem may be detected in the home before driving and such data can be brought into
the vehicle to provide the in-vehicle monitoring system with an initial parameter set to raise the sensitivity of
detection. For example, lower glucose level data or poor sleep quality data measured in the home may be used in the
vehicle for earlier detection of hypoglycemia and frequent micro-sleep occurrence (Table 2). Driver state monitored
in the home and in the vehicle can also make use of cause-effect relationships. For example, failure to take
medicines in the home may increase the possibility of occurrence of hypoglycemic episodes or epileptic seizure
behind the wheel. Such behavioral data measured in the home can also be brought into the vehicle to enhance or
change the in-vehicle intervention strategy (e.g. failure to take anti-seizure medicine locks the ignition). Integration
of on- and off-board driver monitoring will increase accuracy of detection of drivers’ health problems and
strengthen the intervention strategy to avoid crash.
Another example of driving performance data being brought into the off-board network involves continuous
monitoring of the driving environment. Driving includes many complex and parallel cognitive and physical tasks
under certain levels of stress that may magnify certain aspects of a driver’s chronic health problems (e.g.
cardiovascular diseases). Therefore, a driver’s state measured in the vehicle can be brought into a hospital or a smart
home to be integrated with other behavioral and physiological data measured in the home to diagnose the chronic
health state with better accuracy or in an earlier phase.
Ridella 7
IMPLEMENTATION/APPLICATION
Evolution of the current driver state monitoring technology and integration with vehicle-embedded biometric
sensors or wearable sensors with wireless connection to the vehicle (on-board monitoring) should be able to detect
some or all of the acute health declines in real time. Candidates for on-board biometric sensors for each of the
driver’s acute health decline are summarized in Table 2. Most of the biometric sensors for on-board monitoring
shown in Table 2 are still under research or development. Applicability of these sensors to the detection of acute
symptoms needs to be investigated with consideration for the cost and the user acceptance.
If the driver-state monitoring can detect early signs of a decline while the driver is still conscious, a multilayered
intervention strategy using information and assistive vehicle control could be taken. When the driver fails to take
actions to avoid a crash or if the system fails to detect early signs of the decline due to too rapid decline, future
autonomous vehicle technology will be the key to avoiding crashes by bringing the vehicle safely to a stop on the
hard shoulder of the road for the driver who is likely to be unconscious (i.e. Autonomous Emergency Stop System or
AESS, Shunk, 2009, Nissan, 2013). Stopping a vehicle in busy traffic or high speed traffic could induce additional
crashes involving other vehicles. Vehicle-to-vehicle communication to broadcast the emergency signal to
surrounding and following vehicles could be included in the AESS. Accuracy of detection also needs to be high.
Integration of environmental sensors with driver-state monitoring could be considered so that the system activates
the Autonomous Emergency Stop System when the driver is in a health decline and the risk of crash is imminent
(Figure 3).
The contributing factors and co-morbidities associated with the most common injuries also point to interventions
that could benefit the older occupant. While knowing age of a driver or occupant may help in some driving task
assistance, it is less of an indicator of overall health. Sensing occupant bone quality can lead to real-time adaptive
restraint systems that lower the loads on the poorer quality bones of older or less healthy occupants and could help
reduce incidence of rib injuries. Newer technologies such as 4-point belt systems and inflatable seat belts also help
to reduce or distribute chest loading (Ridella, 2012).
Driver-state data measured in the vehicle can be also brought out of the vehicle to the off-board network. Advanced
Automatic Collision Notification System (AACN) is an example that automatically sends notification of a collision
event together with vehicle and driver-state data from the pre-crash phase to a hospital so that the ER will have
sufficient preparation time prior to arrival of the casualties.
Figure 3 . Functional flow of Autonomous Emergency Stop System.
Ridella 8
ISSUES
Privacy, protection and malicious intent
Since the devices and other instrumentation in the vehicle or on the occupant are measuring, monitoring,
transmitting, and/or recording health status information, concerns arise regarding the protection and privacy of this
data. The Health Insurance Portability and Accountability Act of 1996 (HIPAA) sets rights for an individual’s
health information and prescribes rules and limits over who can review and receive that information. This rule
applies to any form of the information whether it is oral, written or electronic. The information that is envisioned to
be collected may not necessarily be covered by HIPAA depending on how it is collected and used. For example, it
may be that the data is only used in a non-identifiable, real-time manner to inform the vehicle that the occupant is
drowsy or has an abnormal condition. In other cases, when linked to driver/occupant identification technology, the
data is personal and possibly subject to HIPAA. It could also be recorded and/or transmitted depending on
application if there is a need to interpret the data in real-time by a health professional or to review the data by law
enforcement at a later date should an event occur. HIPAA rules are very specific about cases such as this, however,
it is likely an interpretation of HIPAA in the vehicle environment may need further review as this technology
develops.
Another consideration is the malicious use of the health information data. Security of the data, whether stored in the
vehicle, in a portable device or other form needs to be protected from other sources that could use or manipulate the
data in harmful ways. For example, a baseline health information data set could be altered either through direct
intervention or electronic methods such that an abnormal event would not be detected by the vehicle or software.
An intervention or denial of service could occur during driving, thus making the system inoperative. There are
endless scenarios that could occur depending on sophistication of the system. Current cybersecurity research
regarding other vehicle communications should include the possibility of how the health and driver monitoring
activities can be protected from mischievous intent.
Performance
Performance of such a system would require extensive testing for reliability, repeatability and reproducibility. In his
short article nearly fifty years ago, one of Waller’s criteria for monitoring for physical impairment was that “few
persons of low accident risk should be falsely categorized in the high accident risk group”. That is, the number of
false-positive identifications should be minimized and conversely, the number of true-positives should be
maximized. It is imperative that these devices are calibrated properly or can be re-calibrated based on manual
feedback from the driver/occupant or perhaps periodically from a software “push” through an application or other
connected technology whether portable or vehicle-based.
Cancelable
Finally, the driver and/or other occupants may reserve the right to cancel or not participate in real-time identification
and monitoring. This may apply when a vehicle is not driven by an owner or designated driver for whom baseline
data exists. Also, rental cars may either not be equipped with such devices or a driver may opt out from
participation.
SUMMARY
The paper discussed the premise that identification of a vehicle driver (and/or occupants) as well as monitoring their
health, mood or behavioral status while driving, may have significant value for safety. It is documented that health
conditions may contribute to increased crash risk and that those with conditions have poorer outcomes should a
crash occur. By monitoring a driver’s health status in real time, possibly comparing to a baseline value, acute
conditions may be detected and a warning , intervention, or other countermeasure may be applied. There is abundant
technology in development as evidenced by manufacturs’ documented research. Also, both traditional and non-
traditional automotive suppliers have been involved in the early vehicle-based technology research, however, the
sports market has dominated the wearable monitoring device development and production. Research projects in
driver behavior, alcohol detection and a host of other technologies in development may lead to new advances in
safety as the population of driver’s age and people are more aware of their own health status.
Ridella 9
REFERENCES
1. American Diabetes Association. (2012). Diabetes and Driving. Diabetes Care, Vol.35, Supplement 1, ppS81-
S86.
2. Badugu,R., Lakowicz,J.R., and Geddes,C.D. (2004). Noninvasive continuous monitoring of physiological
glucose using a monosaccharide-sensing contact lens. Analytical Chemistry, Vol.76, No.3, pp610-618.
3. Baig,M.M., Gholamhosseini,H., and Connolly,M.J. (2013). A comprehensive survey of wearable and wireless
ECG monitoring systems for older adults. Medical & Biological Engineering & Computing Vol. 51, No. 5,
pp485-495.
4. Balanou,E., Gils,M.V., and Vanhala,T. (2013). State-of-the-art of wearable EEG for personalized health
applications. Presentation material at 10
th
InternationalConference on Wearable Micro and Nano Technologies
for Personalized Health. Retrieved February 20, 2014 from
http://phealth2013.eu/uploads/keepitsimple/E.Balanou.pdf
5. Boyle,L.N., Tippin,J., Paul,A., and Rizzo,M. (2008). Driver performance in the moments surrounding a
microsleep. Transportation Research Part F 11, pp126–136.
6. Casson,A.J., Yates,D.C., Smith,S.J.M., Duncan,J.S., and Rodriguez-Villegas,E. (2010). Wearable
Electroencephalography. IEEE engineering in medicine and biology magazine (2010, May/June), pp44-56.
7. FMCSA. (2007) Expert Panel Recommendations: Cardiovascular Disease and Commercial Motor Vehicle
Driver Safety. Evidence Report. Presented to FMCSA, April 2007.
8. FMCSA. (2007). Expert Panel Recommendations: Seizure Disorders and Commercial Motor Vehicle Driver
Safety. Evidence Report. Presented to FMCSA, November, 2007.
9. FMCSA. (2007) Expert Panel Recommendations: Obstructive Sleep Apnea and Commercial Motor Vehicle
Driver Safety. Evidence Report. Presented to FMCSA, November, 2007.
10. FMCSA. (2009) Expert Panel Recommendations: Traumatic Brain Injury and Commercial Motor Vehicle
Driver Safety. Evidence Report. Presented to FMCSA, March, 2009.
11. FMCSA. (2011) Expert Panel Recommendations: 2010 Update Diabetes and and Commercial Motor Vehicle
Driver Safety.Evidence Report. Presented to FMCSA, May ,2007.
12. Ford (2011, May). Ford In-Car Health and Wellness Solutions. Retrieved February 20, 2014 from
http://medicaldesign.com/site-files/medicaldesign.com/files/archive/medicaldesign.com/Medtronic-fact-sheet-
HR.pdf
13. Hanna, R. (2009). The contribution of medical conditions to passenger vehicle crashes. DOT Report, DOT HS
811 219.
14. Hardy, R.N., Watson, J.W., Cook, R., Zioupos, P., Forrester, B., Frampton, R., Page, M., Kennedy, A., Peach,
S., Sproston, P. Development and Assessment of a Bone Scaning Device to Enhance Restraint Performance.
In: Proceedings of the 19
th
International Conference on the Enhanced Safety of Vehicles. Washington, D.C.
15. Hiestand,D.M., Britz,P., Goldman,M. and Phillips,B. (2006). Prevalence of symptoms and risk of sleep apnea
in the US population. CHEST Journal, Vol.130., No.3, pp780-786.
Ridella 10
16. Insurance Institute for Highway Safety. (2014). Status Report, Vol. 49, No. 1.
17. Iowa State University (2005). Smart Home Lab. Retrieved February 20, 2014 from
http://smarthome.cs.iastate.edu/Demos.html
18. Leger,D. (1994). The cost of sleep-related accidents: a report for national commission on sleep disorders
research. Sleep, Vol.17, No.1, pp84-93.
19. Meditech (2011). Finger Pulse Oximeter. Retrieved February 20, 2014 from
http://www.meditech.com.cn/Finger-Pulse-Oximeter
20. Misugi,K., Kanamori,H., and Atsumi,B. (2003). Toyota’s program for universal design in vehicle
development. Universal design for the Toyota “RAUM”, Toyota Motor Corporation, “Raum” press
information
21. Monk,C (2012, August). Driver Alcohol Detection System for Safety (DADSS). Retrieved February 20, 2014
from https://secure.hosting.vt.edu/www.apps.vtti.vt.edu/PDFs/NDRS-2012-
presentations/DADSS%20VTTI%20Naturalistic%20Conference%20Presentation%202012%20v3.pdf
22. NHTSA (1985). Transportation-related sleep research. CAR files study. Report to the Senate Committee on
Appropriations and House Committee on Appropriations. Washngton,DC., US Department of Transportation.
23. NHTSA (2013). Traffic safety for older people - 5-year plan. DOT HS 811 837.
24. NHTSA (2015). Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causation
Survey. DOT HS 812115.
25. Nijsen,T.M.E., Arends,J.B.A.M., Griep,P.A.M., and Cluitmans,P.J.M. (2005). The potential value of three-
dimensional accelerometry for detection of motor seizures in severe epilepsy. Epilepsy & Behavior, Vol. 7,
pp74-84.
26. Nissan (2013, August). Nissan Autonomous Drive: Emergency Stop. Retrieved February 20, 2014 from
http://nissannews.com/en-US/nissan/usa/releases/nissan-announces-unprecedented-autonomous-drive-
benchmarks/videos/nissan-autonomous-drive-emergency-stop
27. Paul,A., Boyle,L.N., Tippin,J., and Rizzo,M. (2005). Variability of driving performance during microsleeps.
Proceeding of the third International driving Symposium on Human Factors in Driver Assessment, Training
and Vehicle Design. pp18-24.
28. Pautan.org (2011, May). Ford Develops Heart Rate Monitoring Car Seat. Retrieved February 20, 2014 from
http://paultan.org/2011/05/26/ford-develops-heart-rate-monitoring-car-seat/
29. Positive ID (2013). Non-Invasive Breath Glucose Detection Device. Retrieved February 20, 2014 from
http://www.positiveidcorp.com/products_breathglucosedetection.html
30. Quick,D. (2009, January). Implantable Sensor Simplifies Blood Pressure Readings. Retrieved February 20,
2014 from http://www.gizmag.com/tiny-sensor-simplifies-blood-pressure-readings/10846/
31. Ridella, S.A., Rupp, J.R., and Poland, K. Age-Related Differences in AIS 3+ Crash Injury Risk, Types,
Causation and Mechanisms. In: Proceedings of the 2012 International IRCOBI Conference on the
Biomechanics of Injury, pp. 43-59, Dublin, Ireland, September 2012.
Ridella 11
32. Ridelust (2011, January). Ferrari Working On Driver Monitoring Technology. Retrieved February 20, 2014
from http://www.ridelust.com/ferrari-working-on-driver-monitoring-technology/
33. Risser,M.R. and Ware,J.C. (1999). Driving simulation with EEG monitoring in nomals and obstructive sleep
apnea patients. Sleep, Vol. 23, No.3, pp393-398.
34. Rizzo,M. (2011). Impaired driving from medical conditions: A 70-year-old man trying to decide if he should
continue driving. JAMA, March 9, 2011, Vol.305, No.10, pp1018-1026.
35. Shunk,C. (2009, June). BMW Working On System That Stops Vehicles During A Medical Emergency.
Retrieved February 20, 2014 from http://www.autoblog.com/2009/06/05/bmw-working-on-system-that-stops-
vehicles-during-a-medical-emerg/
36. Skubic,M., Alexander,G., Poppescu,M., Rantz,M., and Keller,J. (2009). A smart home application to eldercare:
Current status and lessons learned. Technology and Health Care 17, pp183-201.
37. Tregear,S.J., Rizzo,M., Tiller,M., Schoelles,K., Hegmann,K.T., Greenberg,M.I., Phillips,B. and Anderson,G.
(2007). Diabetes and motor vehicle crashes: a systematic evidence-based review and meta-analysis.
Proceedings of the Fourth International Driving Symposium on Human Factors in Driver Assessment,
Training and Vehicle Design. pp343-350.
38. University of Florida. The Gator Tech Smart House (GTSH). Retrieved February 20, 2014 from
http://www.icta.ufl.edu/gt.htm#1
39. University of Rochester Medical Center (2014). University Creates Medical "Smart Home" To Study Future
Health Technology. Retrieved February 20, 2014 from
http://www.urmc.rochester.edu/news/story/index.cfm?id=-103
40. Waller, J.A. (1967). The Contribution of Medical Impairment to Traffic Accidents. In: The Prevention of
Highway Injury, Selzer, M.L., Gikas, P.W. and Huelke, D.F. (eds). , Highway Safety Research Institute, Ann
Arbor, MI.
41. Watson, T. , Krause J., Le, J. and Rao, M.K. (2011). Vehicle Integrated Non-Intrusive Monitoring of Driver
Biological Signals. SAE Paper No. 2011-01-1095., SAE International, Warrendale, PA.
42. Wild,S., Roglic,G., Green,A., Sicree,R. and King,H. (2004). Global prevalence of diabetes; Estimates for the
year 2000 and projections for 2030., Diabetes Care, Vol.27, No.5.
43. Wired. (2012). http://www.wired.com/2012/06/ford-workload/
Ridella 12
Table 2.
Monitoring technologies and application to driver’s state
... In order to monitor the driver's physiological measures in real-time, Ford Motor Company developed a prototype of physiological sensors embedded within the vehicle [62]. The MIT Media Lab also demonstrated that physiological measures such as heart rate and respiration rate can be remotely monitored using a camera installed on the steering wheel. ...
Article
In this article, a dynamic hybrid choice model of the driver stress arising from cognitive workload is developed and applied to a driving simulator experiment conducted with students at the American University of Beirut. A latent or unobserved variable quantifying the state stress over time is integrated with a discrete choice model of red-light violations. The driver state stress is induced by additional cognitive workload and situational factors in an urban driving context and is dependent on a time-invariant agent effect or individual trait. Driving performance (e.g., speed and acceleration) and physiological measures (e.g., heart rate) are used as indicators of the underlying state stress. The driver state stress is found to be significantly affected by road events or situations (e.g., encountering pedestrians, trucks, and traffic light), the level of cognitive workload, the individual propensity for stress, and the mere driving task (e.g., maintaining control of the car). Moreover, results show that there is a pattern of regulatory driving behavior in response to the increasing stress. The developed model is a mathematical conceptualization of the transactional model of the driver stress and can be integrated within in-vehicle systems to detect/predict the driver state and enhance safety and well-being.
Article
Objective: Some lane-keeping assist systems in development and production provide autonomous braking and steering to correct unintentional lane drift but otherwise require drivers to fully control their vehicles. The goal of this study was to quantify the proportion of drivers involved in unintentional lane drift crashes who would be unable to regain control of their vehicles to inform the design of such systems. Methods: The National Highway Traffic Safety Administration's National Motor Vehicle Crash Causation Survey collected in-depth, on-scene data for a nationally representative sample of 5,470 U.S. police-reported passenger vehicle crashes during 2005–07 that occurred between 6 a.m. and midnight and for which emergency medical services were dispatched. The physical states of drivers involved in the 631 lane drift crashes in the sample, which represented 259,034 crashes nationally, were characterized. Results: Thirty-four percent of drivers who crashed because they drifted from their lanes were sleeping or otherwise incapacitated. These drivers would be unlikely to regain full control of their vehicles if an active safety system prevented their initial drift. An additional 13% of these drivers had a non-incapacitating medical issue, blood alcohol concentration (BAC) ≥ 0.08%, or other physical factor that may not allow them to regain full vehicle control. When crashes involved serious or fatal injuries, 42% of drivers who drifted were sleeping or otherwise incapacitated, and an additional 14% were impacted by a non-incapacitating medical issue, BAC ≥ 0.08%, or other physical factor. Conclusions: Designers of active safety systems that provide autonomous lateral control should consider that a substantial proportion of drivers at risk of lane drift crashes are incapacitated. Systems that provide only transient corrective action may not ultimately prevent lane departure crashes for these drivers, and drivers who do avoid lane drift crashes because of these systems may be at high risk of other types of crashes when they attempt to regain control. Active lane-keeping assist systems may need to be combined with in-vehicle driver monitoring to identify incapacitated drivers and safely remove them from the roadway if the systems are to reach their maximum potential benefit.
Article
Full-text available
This study aimed to evaluate the value of measuring microsleeps as an indicator of driving performance impairment in drowsy drivers with sleep disorders. Drivers with sleep disorders such as obstructive sleep apnea/hypopena syndrome (OSAHS) are at increased risk for driving performance errors due to microsleep episodes, which presage sleep onset. To meet this aim, we tested the hypothesis that OSAHS drivers show impaired control over vehicle steering, lane position and velocity during microsleep episodes compared to when they are driving without microsleeps on similar road segments. A microsleep is defined as a 3-14 sec episode during which 4-7 Hz (theta) activity replaces the waking 8-13 Hz (alpha) background rhythm. Microsleep episodes were identified in the electroencephalography (EEG) record by a neurologist certified by the American Board of Sleep Medicine. Twenty-four drivers with OSAHS were tested using simulated driving scenarios. Steering variability, lane position variability, acceleration and velocity measures were assessed in the periods during a microsleep, immediately preceding (pre) microsleep, and immediately following (post) microsleep. In line with our introductory hypothesis, drivers with OSAHS did show significantly greater variation in steering and lane position during the microsleep episodes compared to the periods pre and post microsleep. The results indicate that identification of microsleep episodes can provide a marker for declining vehicle control of drivers with OSAHS.
Article
Full-text available
The electroencephalogram EEG is a classic noninvasive method for measuring a persons brainwaves and is used in a large number of fields: from epilepsy and sleep disorder diagnosis to braincomputer interfaces BCIs. Electrodes are placed on the scalp to detect the microvolt-sized signals that result from synchronized neuronal activity within the brain. Current long-term EEG monitoring is generally either carried out as an inpatient in combination with video recording and long cables to an amplifier and recording unit or is ambulatory. In the latter, the EEG recorder is portable but bulky, and in principle, the subject can go about their normal daily life during the recording.
Article
Full-text available
This study examined if individuals who are at increased risk for drowsy-driving because of obstructive sleep apnea syndrome (OSAS), have impairments in driving performance in the moments during microsleep episodes as opposed to during periods of wakefulness. Twenty-four licensed drivers diagnosed with OSAS based on standard clinical and polysomnographic criteria, participated in an hour-long drive in a high-fidelity driving simulator with synchronous electroencephalographic (EEG) recordings for identification of microsleeps. The drivers showed significant deterioration in vehicle control during the microsleep episodes compared to driving performance in the absence of microsleeps on equivalent segments of roadway. The degree of performance decrement correlated with microsleep duration, particularly on curved roads. Results indicate that driving performance deteriorates during microsleep episodes. Detecting microsleeps in real-time and identifying how these episodes of transition between wakefulness and sleep impair driver performance is relevant to the design and implementation of countermeasures such as drowsy driver detection and alerting systems that use EEG technology.
Article
A vehicle integrated sensing and analysis system has been designed, implemented, and demonstrated to nonintrusively monitor several biological signals of the driver. The biological driver signals measured by the system are the heart electrical signals or pseudo Lead-I electrocardiography (pLI-ECG), the galvanic skin response (GSR) or electrical conductance measured from the driver's fingers to palm, the palm skin temperature, the face skin temperature, and the respiration rate. The pLI-ECG and GSR measurements are made through direct contact of the driver hands with stainless steel electrodes integrated in the steering wheel rim. The temperature measurements are made with non-contacting infrared temperature sensors, also located on the steering wheel. The respiration rate was measured using a flexible thin film piezoelectric sensor affixed to the seatbelt. This system was installed in a test car and on-road driving was performed to assess the effectiveness of the integrated systems in continuously monitoring these biometric parameters over a range of driving conditions. Signals were processed in real time to estimate the short term driver heart rate and breathing rate and all analog signals were stored for additional post processing. Comparisons of the biological signals obtained from this vehicle integrated system were made with those simultaneously obtained from a medical grade measurement system. The signals from the integrated biological monitoring system may be used for assessing driver performance as a function of driver wellness, workload and stress.
Conference Paper
Data from the National Automotive Sampling System‐Crashworthiness Data System (NASS‐CDS) and the Crash Injury Research Engineering Network (CIREN) cases were analyzed to (1) define how occupant age affects crash exposure and the risk of AIS 3+ injury and (2) identify how age affects injury types and causation while considering comorbid factors and gender. Multivariate logistic regression models developed from NASS indicated that age significantly increases the risk of AIS 3+ injury to almost every body region in all crash modes. The largest age effect was observed for the thorax in frontal crashes and was modulated by gender such that older women were at significantly higher risk relative to younger women than were older men relative to younger men. A subset of 1289 CIREN cases, analyzed for injury causation, found occupants under age 25 and over 65 years sustained greater proportions of head and thoracic injuries respectively. For frontal crashes, thoracic injury type changed from soft to bony tissue injury as age increased and comorbid factors were more common for the older occupants. Crash severity was consistently lower for the older age groups and outcomes were worse in the older occupants even for similar crash severity and injury.
Article
Wearable health monitoring is an emerging technology for continuous monitoring of vital signs including the electrocardiogram (ECG). This signal is widely adopted to diagnose and assess major health risks and chronic cardiac diseases. This paper focuses on reviewing wearable ECG monitoring systems in the form of wireless, mobile and remote technologies related to older adults. Furthermore, the efficiency, user acceptability, strategies and recommendations on improving current ECG monitoring systems with an overview of the design and modelling are presented. In this paper, over 120 ECG monitoring systems were reviewed and classified into smart wearable, wireless, mobile ECG monitoring systems with related signal processing algorithms. The results of the review suggest that most research in wearable ECG monitoring systems focus on the older adults and this technology has been adopted in aged care facilitates. Moreover, it is shown that how mobile telemedicine systems have evolved and how advances in wearable wireless textile-based systems could ensure better quality of healthcare delivery. The main drawbacks of deployed ECG monitoring systems including imposed limitations on patients, short battery life, lack of user acceptability and medical professional's feedback, and lack of security and privacy of essential data have been also discussed.
Article
Some medical disorders can impair performance, increasing the risk of driving safety errors that can lead to vehicle crashes. The causal pathway often involves a concatenation of factors or events, some of which can be prevented or controlled. Effective interventions can operate before, during, or after a crash occurs at the levels of driver capacity, vehicle and road design, and public policy. A variety of systemic, neurological, psychiatric, and developmental disorders put drivers at potential increased risk of a car crash in the short or long term. Medical diagnosis and age alone are usually insufficient criteria for determining fitness to drive. Strategies are needed for determining what types and levels of reduced function provide a threshold for disqualification in drivers with medical disorders. Evidence of decreased mileage, self-restriction to driving in certain situations, collisions, moving violations, aggressive driving, sleepiness, alcohol abuse, metabolic disorders, and multiple medications may trigger considerations of driver safety. A general framework for evaluating driver fitness relies on a functional evaluation of multiple domains (cognitive, motor, perceptual, and psychiatric) that are important for safe driving and can be applied across many disorders, including conditions that have rarely been studied with respect to driving, and in patients with multiple conditions and medications. Neurocognitive tests, driving simulation, and road tests provide complementary sources of evidence to evaluate driver safety. No single test is sufficient to determine who should drive and who should not.