Real-Time Monitoring of Driver Distraction:

State-of-the-art and Future Insights

Eva Michelaraki1*, Christos Katrakazas1, Susanne Kaiser2, Tom Brijs3, George Yannis1
1
National Technical University of Athens, Department of Transportation Planning and Engineering,
5 Heroon Polytechniou str., GR-15773, Athens, Greece
2
KFV, Austrian Road Safety Board, Schleiergasse 18, 1100 Wien, Austria
3
UHasselt, School for Transportation Sciences, Transportation Research Institute (IMOB), Agoralaan, 3590 -
Diepenbeek, Belgium

Abstract

Human-related factors, especially driver distraction and inattention, are major contributors to a large number of
serious road crashes. It is evident that distraction reduces to a great extent driver perception levels as well as
decision making capability, which negatively affects driver's ability to control the vehicle. An effective way to
reduce these kinds of crashes would be through monitoring drivers' mental state or driving behaviour and alerting
them when they are in a distracted state. In recent years, several inexpensive and effective detection systems have
been developed in order to mitigate driver inattention. This study aims to critically review and assess the state-of-
the-art in driver attention measuring, as well as the corresponding technologies for risk assessment and mitigation,
as part of the i-DREAMS project. A thorough literature review was carried out in order to compare and contrast
technologies that can be used to detect, monitor or measure driver's distraction or attention. In most of the identified
studies, driver distraction was measured with respect to its impact to driver behaviour. Real-time eye tracking
systems, cardiac sensors on steering wheels, smartphone applications and cameras were found to be the most
frequent devices to monitor and detect driver distraction. On the other hand, less frequent and effective approaches
included electrodes, hand magnetic rings and glasses.

Keywords: Distraction, Attention, State-of-the-art Technology, Inattention Monitoring System, Driver State
Monitoring.

1
Corresponding author. Tel.: +30-210-772-1265;
E-mail address: [email protected]

1
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

1. Introduction
Approximately, 1.25 million people die every year on roads worldwide, with millions more sustaining serious
injuries and living with long-term adverse health consequences [1]. Globally, road crashes are one of the leading
causes of death, especially among young people, as well as the number one cause of death among those aged 15–
29 years [2]. Currently, road crashes are estimated to be the ninth leading cause of death across all age groups, and
are also predicted to become the seventh dominant cause of death by 2030 [1].

Several human factors have been identified which affect the likelihood of a road traffic crash or a serious injury,
but among them, driver distraction or inattention are some of the major contributors demonstrating the increased
risk of road traffic fatalities and injuries [3, 4]. It is worth mentioning that limiting the exposure to these risk
indicators is essential and critical to the success of efforts in order to reduce traffic injuries and therefore promote
road safety.

Specifically, driver distraction (in-vehicle or external) represents an important factor of driver state with negative
impact on road safety and is a major cause of vehicle crashes worldwide with an increasing importance [5]. At the
same time, technological developments make massive and detailed operator performance data easily available, via
new in-vehicle sensors that capture detailed driving style. This creates new opportunities for the detection and
design of customized interventions to mitigate the risks, increase awareness and upgrade driver performance,
constantly and dynamically [6]. The optimal exploitation of these opportunities is the challenge that i-DREAMS
faces.

The overall objective of the European H2020 i-DREAMS2 project is to define, develop, test and validate a context-
aware safety envelope for driving in a ‘Safety Tolerance Zone’ (STZ), with a smart Driver, Vehicle & Environment
Assessment and Monitoring System. Taking into account, on the one hand, driver background factors and real-
time risk indicators, and on the other hand, driving task complexity indicators, a continuous real-time assessment
will be created to monitor and determine if a driver is within acceptable boundaries of safe operation (i.e. STZ).
Testing and validation will be applied to car, bus and truck drivers as well as to tram and train drivers.

Within, the above framework, the aim of the work documented in this paper is to review and assess state-of-the-
art in-vehicle approaches and technologies as well as the various driver recording tools to monitor the driver's
distraction and inattention. To achieve this objective, a comprehensive literature search (scientific as well as grey
literature) was conducted. Identified measurement methods and associated technologies were assessed based on
pre-defined criteria such as intrusiveness and effectiveness among others. The review was conducted from a
transportation mode perspective, beginning with car technologies which were covered most extensively in
literature. Following this, the transferability of the results to another modes (i.e. buses, trucks and trains/trams)
was assessed and if necessary, a dedicated further search for a certain mode was carried out.

The paper is structured as follows. In the beginning, the overall objective of the i-DREAMS project as well as the
aim of this research is provided. Subsequently, the theoretical background of driver distraction definition and
corresponding indicators is given. This is followed by a section, in which, several definitions with regards to the
terms of driver distraction and inattention are presented and the types of driver distraction are also analyzed.
Moreover, the methodological approach of the current research is presented. An extended literature review was
carried out regarding all available state-of-the-art technologies of assessing driver distraction. In the next step, the
results of technologies and systems that has been identified for the real-time monitoring of driver inattention are
presented. Finally, overall conclusions for the continuous monitoring of driver distraction are highlighted in order
to assist researchers and practitioners.

2. Defining distraction and corresponding indicators

The use of different, and sometimes inconsistent, definitions of driver distraction can create a number of problems
for researchers and road safety stakeholders. Definitions currently utilized in the literature are either from relevant
scientific studies, or official bodies and organisations. For example, the International Standards Organisation (ISO)
defines driver distraction as being "attention given to a non-driving related activity, typically to the detriment of
driving performance" [7]. In addition, Young et al. [8] considered that distraction is defined as "the presence of an
event, activity, object or person within or outside the vehicle which compelled or tended to induce the driver’s

2 Further general project information can be found on the website: https://idreamsproject.eu

2
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

shifting attention away from the driving task", while Stevens and Minton [9] proposed driver distraction as
"physical events, actions or conditions, in or on the vehicle that divert attention from driving". According to Streff
[10], distraction involves "a shift in attention away from stimuli critical to safe driving toward stimuli that are not
related to safe driving" and Regan and Strayer [11] claimed that distraction can be defined as "a diversion of
attention away from activities critical for safe driving toward a competing activity".

As there is not a consistent definition for driver distraction and driver inattention across studies, the comparison
among them, may be difficult or sometimes impossible. Even seemingly, similar works sometimes investigate
slightly different concepts or measure different outcomes. It is worth mentioning that inconsistent definitions may
also lead to disparate results of road crash data and therefore, to contrasting estimates or assumptions of the role
of distraction in road accidents. Consequently, these concerns highlight the need to develop a common, generally
well-accepted definition of driver distraction.

For that purpose, driver distraction can be defined as "a diversion of attention from driving, because the driver is
temporarily focusing on another event, task, object or person which is not related to driving" [5]. As a result, the
driver's awareness, decision making ability as well as driving performance are reduced, leading to an increased
risk of corrective actions, near-crashes or crashes. Following the definition above, the current study focuses on
identifying the ways in which distraction and inattention can be monitored during trips and less attention is given
to the relationship between driver distraction and road safety. For instance, Papantoniou et al. [4] provided two
very interesting approaches with a review of driving performance parameters critical for distracted driving with
regards to road safety [12].

As real-time measurement of physiological and behavioural indicators is crucial (especially for the i-DREAMS
concept), the most important indicators will be introduced below with definitions and descriptions. In general,
physiological measures are devoted primarily to continuous measurement of the physical responses of the body,
for example, heart rate or heart rate variability. The most reliable and sensitive physiological measures include eye
movements, such as eye blink rate, blink duration, fixations, saccades and interval of closure as well as head
movements, such as rotation and orientation. A range of driver distraction measures, as well as their indicators that
have been used to evaluate the impact of distraction on driving performance is provided in Table 1, including
behavioural (i.e. longitudinal control, lateral control, reaction time, gap acceptance), as well as physiological
measurements (i.e. eye and head movements).

Table 1: A range of driver distraction measures with their indicators

Driver distraction measures Indicators
Longitudinal control [13, 14] speed, headway
Lateral control [15, 16] lateral position, steering wheel control, standard deviation of steering wheel angle
Reaction time [17, 18] perception response time (PRT), brake response time (BRT), time-to-collision (TTC)
Gap acceptance [19] number of collisions, gaps accepted
Eye movements [20-22] glances, saccades, fixations, blinks, gaze direction, eyes-off-road-time,
electrooculography (EOG), percentage of eyelid closure time (PERCLOS),
percentage of time spent not looking ahead (PERLOOK)
Head movements [23-26] rotation, orientation, pose

Driver distraction is a multidimensional phenomenon and there is not a unique driving performance measure which
is able to capture all effects of distraction. The large number of driver distraction measurements, presented in Table
1, indicates that the decision regarding which measure or set of measures is used should be guided by the specific
research question [5]. In addition, visual distraction has a greater effect on lateral control measures, whereas
cognitive distraction effects more visual scanning behaviour. Among all trackable parameters, longitudinal and
lateral control measurements, surrogate safety measures such as reaction time or gap acceptance, and eye or head
measures are deemed to be the most crucial to identify driver distraction. However, the diversity in the measures
used, in combination with the diversity in the design of the experiments (i.e. road and traffic factors), often
complicated the synthesis of the results, especially for less commonly examined distraction factors.

3. Methodology
In order to review and assess the state-of-the-art attention and distraction measurement techniques, a systematic
search of relevant scientific and grey literature was carried out. Although there was a range of studies investigating
the impact of attention and distraction in the context of road safety, this literature search and review explicitly
focused on research relating to objectively measuring and detecting driver distraction and inattention during trips,

3
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

preferably in real-time driving conditions. The key terms were then entered into the databases, with the following
inclusion criteria:

 Published between 2000-2020

 Search term included in title, abstract or key words
 Language as English
 Document type as journal or review
 Source type journals

The search was conducted in the databases ScienceDirect, ResearchGate, Scopus, PubMed and Google Scholar.
Publications were deduplicated, screened by title (624 publications) and then by abstract. Relevant literature was
documented and summarized. The limitation was set to publications after 2000 and only publications from peer-
reviewed English language journals were considered for inclusion. Additional key references were also added.
Eventually, 29 publications were screened thoroughly. The literature predominantly concerned car driving,
however, the extent of the transferability of the findings to the other i-DREAMS modes (i.e. truck, bus, train and
tram), was discussed.

4. Results
The results of the literature review revealed a variety of different sensors and systems that have been selected to
detect driver distraction. The most prominent technologies that were applied, not only in the academic field, but
also commercially applications, were driver facing cameras (Delphi Electronics, Optalert, Mobileye), eye tracker
systems (EyeAlert, SensoMotoric Instruments, Seeing Machines, Smart Eye, Phasya) or glasses (Tobi eye-tracking
glasses), smartphone applications (CarSafe), wearable devices (BioRadio, FlexComp, Shimmer 3, Empatica E4
Wristband) and steering angle sensors (Cardio Wheel, Texas Instruments Biometric Steering Wheel). The research
literature documents two types of measures associated with periods of distraction or inattention: physiological and
behavioural indicators.

4.1. Physiological indicators

In the past few years, many researchers have been working on the development of safety monitoring technologies
using different techniques. To begin with, Toyota and Lexus' Driver Attention Monitor have been conceived to
detect driver attentiveness, using infrared sensors and cameras monitoring the driver’s face [27]. This technology
is able to identify the driver's face orientation and facial expressions. With regards to the latter, previous works on
detecting driver behaviour proved that facial movements provide useful information associated with secondary
tasks, such as talking [28]. In particular, features related to brow motion and eye lids movements can be used to
capture signaling cognitive load [29]. Moreover, the system found to be non-intrusive solution for real-time
distraction monitoring, providing flashing lights and warning sounds. If no action is taken, the vehicle applies the
brakes (a warning alarm sounds followed by a brief automatic application of the braking system). Fernández et al.
[30] proposed the EyeAlert system as an ideal technology which focuses entirely on the driver’s alertness levels
or distraction from the road ahead. When the infrared camera or sensors monitor driver's eye closure rate, or blink
duration and unsafe patterns are identified, an audible alarm is sounded. According to the available product
information, the portable device focuses on the driver's inattention to the road ahead and it was revealed to be an
effective technology which works regardless of weather or roadway conditions such as fog, snow or rain.

Delphi Electronics, developed a real-time vision-based camera Driver Status Monitor [31]. By detecting drivers'
facial characteristics, this technology analyzed eye-closures and head pose in order to infer their distraction and
inattention. In addition, the system found to be an effective and non-intrusive solution which provided real-time
warnings and notifications and prevent drivers from being too distracted with non-driving tasks. Furthermore,
SensoMotoric Instruments (SMI) provided an eye-tracking system which measured gaze direction, head pose and
orientation, PERCLOS, eyelid closure and blink, as well as pupil diameter and position [32]. It was revealed that
SMI’s InSight computer-based system was an effective way to monitor driver distraction, however user calibration
was necessary. Furthermore, Seeing Machines is an effective and non-intrusive face and eye-tracking system,
monitoring the movements of a person’s eyes, face, head, or facial expressions and distraction events in real-time
through in-cab sensors and cameras [33].

In the same way, Smart Eye is an eye-tracking system measuring eye fixation pattern, smooth pursuit of eye
movement, blink rate and eye lid control through cameras on dashboard [34]. An exit survey was conducted by
Kumar et al. [35] revealed that Smart Eye device is a user friendly, cost-effective and easily accessible oculomotor

4
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

monitoring tool and it did not appear to be an intrusive solution. As all of the participants claimed that they did not
face any difficulty in understanding the visual pursuit task, it was an effective technology for detecting driver
inattention. In addition, the early warning distraction technology of Optalert, was found to be a non-intrusive
solution, as drivers did not need to wear or do anything [36]. The product is designed to detect various signs, such
as facial features, changes in the pattern of eye and head movements and skin conductance. A very interesting
finding by Corbett [37] revealed that drivers indicated that Optalert would be the preferred and effective option as
a research tool related to distraction. Results also confirmed that the system leads to improved safety and warns
the drivers when their distraction exceeds predetermined levels through visual and audio alerts.

Cardio Wheel, an Advanced Driver Assistance System found to be an effective and unintrusive solution that
acquired the electrocardiogram (ECG) from the driver’s hands via sensors on the steering wheel to continuously
detect distraction [38]. One of the most important advantages of this technology is that it can be integrated with
certain third-party systems, such as Mobileye and GeoTab, providing complete fleet management solutions for
enhanced road safety. Furthermore, Texas Instruments Biometric Steering Wheel is a non-intrusive technology for
measuring driver distraction but no information was found about the validity of this technology [39]. Texas
Instruments proved a concept of how biometric sensors mounted on a steering wheel can be used to obtain
important information from a driver in real-time, on condition that simple hand contact is required [40]. This
product combines modern solid-state technology with low-power processing ability and wireless communication
to detect respiration rate, pulse rate as well as ECG-based heart rate from a standalone system. According to the
available technology information, it was found that it cannot be used in real-time conditions and it is not available
for sale. However, it can be only available for testing in a simulator environment.

BioRadio by Great Lakes NeuroTechnologies is a wearable device that acquires physiological data for detecting
driver distraction through electrodes attached to fingers [41]. It is easy to set up and operate and the wearable
wireless physiology monitor can stream data to a computer via Bluetooth or save it to memory for mobile
monitoring. Results indicated that the wireless connectivity and the way it records signals was seamless and noise-
free. Furthermore, FlexComp from Thought Technology is a wearable device using electrodes attached to fingers
[42]. Sensor locations can be set on non-intrusive locations on the body such as on forearm, as there are less
intrusive than on fingers and less sensitive to movement. This technology is mainly used for biofeedback and the
transferability to the driving context is not clear. As the above wearable devices are wireless, portable and easy to
use technologies, they are a suitable solution for clinical research and teaching, but they can be utilized only with
assistance of project staff in a simulator study and they are not appropriate for on-road tests.

In addition, Shimmer 3 provides a wearable device using wireless and robust body worn sensors [43]. Specifically,
photoplethysmography (PPG) ear clip using electrodes position on chest or arms, while a GSR unit, EMG3 using
electrodes attached to two fingers. Shimmer 3 is a good solution for a simulator study as there are non-invasive
and low-cost sensors, suitable for in car use. However, results have to be further validated with a larger sample as
there is was only small sample size examined. Furthermore, Empatica E4 Wristband is a wearable device, equipped
with sensors that offers real-time high-quality physiological data [44]. It was found to be an effective, easy to use
and non-intrusive technology for the identification of driver distraction. The system's battery runs 48 hours and an
internal memory allows to record for up to 60 hours of data. As a result, Empatica E4 is an ideal solution for
longitudinal studies.

Hand sensors such as a hand magnetic rings or magnetic eyeglasses clips were found to be less frequent approaches
in order to monitor driver distraction [24]. Tobi eye-tracking glasses are less effective for monitoring driver
distraction as the calibration of eye tracker might be time-consuming [45]. Eye-tracking glasses are intrusive as
drivers are required to wear them during driving. Results indicated that this technology was not suitable for on-
road trials.

4.2. Behavioural indicators

Mobileye solution is a forward facing camera, which alerts drivers when an imminent rear-end collision is looming,
helps to keep a safe following distance, warns then about unintentional lane departures, and provides indications
about the detected speed limit signs. It was found to be an effective and non-intrusive solution for monitoring the
adverse consequences of driver distraction, promoting road safety. Moreover, it should be noted that smartphones,
with their embedded sensors, such as gyroscopes, accelerometers and magnetometers, were found to be promising
tools for monitoring driving behaviour effects of distraction [46]. Smartphone applications which can provide
measures such as lateral and longitudinal acceleration, can be utilized for surrogate safety measures capturing
observed distraction and inattention. For instance, You et al. [47] presented a driver safety application, called

5
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

CarSafe, which detects drivers to dangerous driving conditions as well as inattentive driving and alters the drivers
accordingly.

It should be mentioned that since smartphones are portable devices, they are more related to the person who carries
them, than to the car. This implies that these devices are not directly linked to the car structure, well-fitting many
vehicle types. Smartphone solutions are increasing in vehicle telematics because they are scalable, upgradable and
low cost. Also, they can provide instantaneous driver feedback and have many embedded sensors. Issues that have
to be considered are the low quality of the sensors, which are not primarily selected for vehicular measurements.
Moreover, smartphones are not fixed, leading to issues as regarding relative orientation, driver/passenger
recognition and GNSS coverage.

5. Discussion
In order to monitor driver distraction and inattention, several hardware and software systems and technologies
were examined. One of the main conclusions that can be drawn is that the most frequently utilized method for the
continuous driver monitoring found to be the use of physiological indicators. Eye movements such as the number
and duration of eye fixations as well as ECG measures and head movements are indicated to be the most reliable
ones. It is worth mentioning that the majority of the studies reviewed, were conducted and tested mostly in driving
simulated environments with limited studies using open field driving experiments with real road conditions within
a specific transport mode. This result is plausible due to the danger of testing inattention on road driving
environments, given the ethical constraints that come with inducing distraction. Also, a manipulation check is
easier to conduct in the controlled environment of a simulator. However, the results obtained in a driving simulator
study may be applied to a real traffic environment. In addition, in driving simulators, there was not found a
particular technology, device or navigation system which was directly connected into the vehicle for distraction
monitoring. For instance, no product was able to discriminate between cars' or trains' interior. Consequently, all
methods that were developed from driving simulator experiments in order to measure distraction and inattention,
were easily transferable to different transport modes. Table 3 provides an overview of devices and technical
equipment used in the reviewed studies on measuring distraction and inattention.

Table 3. Overview of devices and technical equipment used in the reviewed studies on measuring
distraction and inattention
Equipment, On Road
Product/ Simulator Overall assessment and considerations for i-
measurement Intrusiveness Test Indicators
Technology (Yes/No) DREAMS project
method (Yes/No)
Seeing machines In-cab sensor, Contact free Yes Yes Face and eye + Used in truck fleets
cameras facing tracking indicators + Planned to be used on UK rail network
forward and driver + Already implemented in trams following tram crash
+ Can be designed to issue alerts
+ Established product
+ Combine multiple camera sensors to detect a wider range
of movements
- Need installation and training on use/analysis
- No clear results on time headway
Optalert Video cameras on Contact free Yes Yes Vision based: eye + Issues early warnings
dashboard, tracking, facial + Driver does not need to wear or do anything
steering wheel features, amplitude + Established product
and velocity ratio of - Need installation and training on use/analysis
blinks - Licensable software

Cardio wheel Sensors on Low Yes Yes ECG, HRV + Dashboard for fleets of vehicles
steering wheel + Can be integrated with certain third-party systems
+ Non-intrusive
- Requires contact of both hands to steering wheel
- Requires custom steering wheel
Smart eye Eye tracking Contact free Yes Yes Vision based: eye, + Developed for automotive industry
cameras on face and head + Established product
dashboard tracking + Non-intrusive
+ Driver does not need to wear or do anything
- Need installation and training on use/analysis
- Sensitivity of the system

Texas Instruments Sensors on Low Yes Yes ECG heart rate, + Fully assembled board developed for testing and
Biometric Steering steering wheel, pulse rate, distraction validation
Wheel measuring pulse, respiration rate + Non-intrusive
respiration - Requires contact of both hands to steering wheel
- Unsure of validation
- Not available for sale

6
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

Equipment, On Road
Product/ Simulator Overall assessment and considerations for i-
measurement Intrusiveness Test Indicators
Technology (Yes/No) DREAMS project
method (Yes/No)
BioRadio by Great Electrodes Medium Yes No ECD, EMG + Wireless connectivity
+ Seamless and noise-free recorded signals
Lakes attached to finger
+ Used for clinical research
NeuroTechnologies (in study: placed
- Signals with accuracy of only 70%
on steering wheel) - Use only with assistance of project team
Empatica E4 Wristband with Low Yes Yes EDA sensor + Easy to use technology
+ Battery runs 48 hours
Wristband sensors
+ Internal memory with up to 60-hour recorded data
+ Additional equipment: 3-axis Accelometor to capture
motion-based activity, event-mark button
- No details on HR parameters provided
FlexComp from Electrodes Medium Yes No HRV, EDA + Sensor locations can be set on non-intrusive locations on
Thought Technology the body
attached to fingers
+ High quality signals
+ Easy to use technology
+ Use fiber optic for real-time monitoring
- Mainly used for biofeedback
- Transferability to driving not clear
- License restriction is enforced through limitations on the
software's functionality
Shimmer 3, including Electrodes Medium Yes No ECG + Electrodes are placed on participant fixed for each trial
PPG ear clip + Can be used with other devices
positioned on chest
or arms - Used in laboratory research

Shimmer 3 GSR unit, Electrodes Medium Yes No EMG + Electrodes are placed on participant each trial
+ Non-invasive sensors
EMG3 attached to two
+ Low-cost sensors
fingers
+ Suitable for in car use
- Results have to further validated with a larger sample
- Small sample size
- Used in laboratory research
- Cannot be used together with Shimmer 3, which measures
ECG

Tobi eye-tracking Eye-tracking light Medium Yes No Vision based: mean + Very detailed documentation of operationalization, design
and procedure
glasses glasses fixation time
- Calibration of eye tracker might be time-consuming
- Eye-tracking glasses medium-intrusive
- Not suitable for on-road real-time trials
EyeAlert Eye tracking Low Yes Yes Vision based: eye, + Visual feedback and auditory/voice alarms
+ Works in all weather and road conditions
cameras on face and head
+ Small and portable device
dashboard tracking
+ Mounts easily on the dashboard
- Sensitivity of the system
Delphi Electronics Vision-based Low Yes Yes Eye-closures and + Real-time warnings and notifications
+ Non-intrusive
Driver Status single camera head pose
+ Increases road safety
Monitor
+ Offers the most direct indication of early distraction
- Sensitivity of the system
SensoMotoric eye tracking Low Yes Yes Gaze direction, head + Quality product
+ Usability
Instruments system, computer position, eyelid
+ Validation to obtain solutions that are truly focused on
vision-based closure and
resolving an unmet need
PERCLOS - User calibration is necessary
Toyota and Lexus' Eye tracking Low Yes Yes face orientation and + Provides flashing lights and warning sounds
+ Prevents frontal collisions, unintended lane departures and
Driver Attention cameras on facial expressions
night-time accidents
Monitor dashboard, sensors
+ Intuitive features for enhancement driver's awareness of
surroundings
- System may in some cases not operate properly due to a
variety of road/vehicle/weather conditions
- People and obstacles that show in the monitor differ from
the actual position and distance
CarSafe application Smartphone Contact free Yes Yes Lateral and + Portable, scalable, upgradable and cheap device
+ Provides instantaneous driver feedbacks
application, front- longitudinal
- Application is only available for Android phones
facing cameras, acceleration
- Low quality of the sensors
embedded sensors - Are not fixed, leading to issues as regarding relative
orientation, driver recognition and GNSS coverage

5.1. Limitations

A few limitations can be arguably found in the current literature with regards to the review and overall assessment
of state-of-the-art real-time technologies for monitoring driver distraction. First of all, one limitation lies in the
nature of the design of the works themselves. As mentioned above, the majority of the researches and systems
investigated, were tested mostly in simulated environments instead of real driving ones, probably due to the danger
of testing inattention in real driving conditions as well as due to the problems of vision systems working in outdoor
environments (i.e. lighting changes, sudden movements). Driver's physiological reactions or movements,

7
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

distraction or inattention may be different in a driver simulator from those in real conditions. Also, simulations are
known to underrepresent on-road conditions, making them less representative and effective solutions, while drivers
may not face the driving process seriously, or sometimes an extra discomfort may be added to participants, usually
caused by simulator sickness.

Additionally, it is worth mentioning that most studies and devices examined, were used in cars and can be
presumed to have been selected for their customizability. As there was not found any study concerning professional
drivers in heavier vehicles, therefore, the findings may not be as useful to other transport modes. It should be also
noted that eye tracking systems or devices operating with cameras in order to monitor the driver's eye movements
or orientation may not function properly when the driver is wearing sunglasses. Furthermore, the functioning may
be dubiously in extremely bright or poor light conditions as well as may be not effective for people with very dark
skin and atypical facial shape. Technologies or wearables should be compatible with safety, prescription or
sunglasses. There was also a noticeable lack of studies that focus on the indirect effects of driver distraction and
inattention. For instance, the vast majority of eye tracking systems and cameras monitored the driver without
detecting the road environment, such as a pedestrian crossing the street, road layout, traffic conditions, time of the
day or weather (i.e. fog, snow, rain). Finally, it is worth noting that some technologies seemed very easy to use
and handle with, but non-professional drivers and customers may be not able to buy them, due a high cost.

5.2. Future research directions

Undoubtedly, future scope of research would be to examine different state-of-the-art systems, products and
technologies testing in real-time on road conditions, as the field validation would increase the reliability of the
findings. Taking into account that simulators may provide contradictory, inconsistent and conflicting results and
produce invalid research outcomes, systems should be validated in real conditions. Lack of detail, general
limitations and underlying biases have to be relied upon to reach a conclusion and researchers must resort to
identify which technologies for monitoring driver distraction or inattention are suitable to real-time naturalistic
driving experiment. Due to the circumstance that each monitoring method or technique described above is not the
one and only standard in research, a thorough testing in real-time conditions for different transport modes is
indispensable. As all studies were conducted with devices resembling car interiors, therefore it could be beneficial
to examine technologies that are able to detect buses', trucks', trains' or trams' driver distraction.

In future studies, technologies for monitoring driver distraction in real-time should be tested with bigger sample
sizes and for longer periods of time. In this way, fuzzy knowledge base will be easier to be generalized. For
instance, it would be a good idea to examine vision systems, especially in drivers wearing glasses in order to solve
the problems for daytime operation. A combination of sensors, such as a steering wheel and a lateral position
sensor in addition to the visual information would be beneficial to achieve a correct detection of driver distraction.

Finally, as the majority of measurements concerning distraction are frequently intrusive, in-vehicle personalization
[48] or the correlation of driving behaviour data with visual measurements through advanced statistical or machine
learning approaches could become beneficial. Such advanced methodologies have been found to be advantageous
in many road safety aspects e.g. activity recognition [49], real-time conflict prediction [50], and its practical and
more generalized application on driver distraction could bring new insights in how to tackle the problem for safe
driving [51].

6. Conclusions
The objective of this study was to review and assess state-of-the-art technologies and systems to monitor driver
distraction and inattention. In addition, a selection of driver inattention factors including measurement methods
were summarized and driver distraction indicators were reviewed.

Technologies and equipment used in the reviewed studies measuring inattention and distraction were separately
reviewed and assessed in terms of intrusiveness and effectiveness and overall applicability for the i-DREAMS
project purposes. An assessment of available technology was provided, focusing on the theoretical suitability of
single devices or technologies for measuring the driver state constructs in question and the applicability in two
settings 'simulator' and 'on-road trial'. Intrusiveness was the main reason for a negative assessment of a device for
the on-road setting as well as prioritization of positive assessed devices was made in terms of effectiveness.

Currently, there is no standard procedure for measuring the driver’s distraction, with a plethora of methods,
indicators and algorithms, each with strengths and drawbacks. In most cases, driver distraction was measured in
terms of its impact to driver attention and driver behavior [4]. The best studied forms of distraction were not only

8
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

visual but also cognitive distraction. Non-intrusive technologies were strongly preferred for measuring inattention,
and vision-based systems have appeared to be the most attractive one not only for drivers but also for researchers.
In particular, attention monitoring systems, including real-time head, gaze and eye tracking systems, sensors on
steering wheels, smartphone applications, wearables and dashboard cameras were found to be the most frequent
devices to monitor and detect the driver's distraction, with head position, viewing and scanning patterns and
PERCLOS being the most reliable indicators. On the other hand, less frequent approaches included hand magnetic
rings and glasses. Non-intrusive methods were strongly preferred for monitoring distraction, and vision-based
systems, providing physiological indicators, have appeared to be attractive for drivers. Nevertheless,
complementing a specific technology with an Electrodermal Activity (EDA) measuring device, such as a
Wristband or a (thermal) camera facing the participant will be beneficial, as the complementary method may
provide evidence for validity.

Regardless of the measurement methods and their quality, practical considerations for implementations in i-
DREAMS should be noted. The vast majority of reviewed literature and information concerned car driving. An
assessment was conducted to see to what extent the conclusions were transferable to other modes. It was revealed
that most of methods, technological devices and systems mentioned above, which measure driver distraction or
attention, can be easily transferred to all transport modes and no indication was found that contradict the
assumption that the identified methods can be transferred from the context car to the other i-DREAMS modes:
trucks, buses, trains and trams.

Wearable devices, such as eye tracking glasses were found to work only with the assistance of project staff in a
simulator study and they were not available for on-road testing. It should be clearly mentioned that the impact on
the naturalistic driving character has to be considered when asking the participants to wear a device whenever they
drive. For instance, when using cameras facing the participant, GDPR is to be considered carefully. Hence, with
the exception of wearables, it can be concluded that attention monitoring systems are easily transferrable to all
four modes of i-DREAMS. This could be very important for the project, providing flexibility, meaning that the
system does not need to be redesigned for each mode of transport.

Systems aimed at increasing driver safety to be effective, an as accurate as possible risk monitoring instrument is
required. This issue will constitute the project’s first pillar (i.e. real-time risk monitoring). Moreover, impact on
driver safety can be expected to be higher, if proposed technologies in some way combine the local perspective
(i.e. in-vehicle assistance with instant impact on driving) with the general perspective (i.e. longer-term support for
a gradual change process in the vehicle operator). The development, implementation and testing of the best and
most suitable technological solution (i.e. the i-DREAMS platform) could bring together these functionalities.

Within the i-DREAMS framework, the conclusions drawn from this study serve as the base for selecting
appropriate measuring systems and devices for the future project work and for building the theoretical and
mathematical model which are the backbone of the development of the i-DREAMS platform. Constructs to be
measured are the driver's cognitive and affectional state (mental state) in terms of attention and distraction as well
as more stable characteristics which are known to impact safe driving. Another outcome will be a research database
with rich information of simulator and on-road drives of hundreds of participants. Since the database aims to
facilitate future research, it can be argued that the more known about the test subjects, the better. However, this is
a question of time, resources and also reasonableness towards participants volunteering to support the i-DREAMS
research.

Acknowledgments
The research was funded by the European Union's Horizon 2020 i-DREAMS project (Project Number: 814761)
funded by European Commission under the MG-2-1-2018 Research and Innovation Action (RIA).

References
1. World Health Organization. (2015). Global status report on road safety 2015. World Health Organization.
https://apps.who.int/iris/handle/10665/189242.
2. World Health Organization. (2008). The global burden of disease: 2004 update. World Health Organization.
https://apps.who.int/iris/handle/10665/43942.
3. Brookhuis, K. A., & De Waard, D. (2010). Monitoring drivers’ mental workload in driving simulators using physiological
measures. Accident Analysis & Prevention, 42(3), 898-903.
4. Papantoniou, P., Papadimitriou, E., & Yannis, G. (2017). Review of driving performance parameters critical for distracted
driving research. Transportation research procedia, 25, 1796-1805.

9
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

5. Regan, M. A., Lee, J. D., & Young, K. L. (2008). What drives distraction? Distraction as a breakdown of multilevel
control. 41-56. CRC Press, Boca Raton, Fla, USA.
6. Horrey, W. J., Lesch, M. F., Dainoff, M. J., Robertson, M. M., & Noy, Y. I. (2012). On-board safety monitoring systems
for driving: review, knowledge gaps, and framework. Journal of safety research, 43(1), 49-58.
7. International Standards for Organization. (2004). Road vehicles – Ergonomic aspects of transport information and control
systems – Occlusion method to assess visual distraction due to the use of in-vehicle systems.
8. Young, K., Regan, M., & Hammer, M. (2007). Driver distraction: A review of the literature. Distracted driving, 2007,
379-405.
9. Stevens, A., & Minton, R. (2001). In-vehicle distraction and fatal accidents in England and Wales. Accident Analysis &
Prevention, 33(4), 539-545.
10. Streff, F. M. (2000). Driver distraction, aggression, and fatigue: a synthesis of the literature and guidelines for Michigan
planning.
11. Regan, M. A., & Strayer, D. L. (2014). Towards an understanding of driver inattention: taxonomy and theory. Annals of
advances in automotive medicine, 58, 5.
12. Kaiser, S., Eichhorn, A., Aigner-Breuss, E., Pracherstorfer, C., Katrakazas, C., Michelaraki, E., Yannis, G., Pilkington-
Cheney, F., Talbot, R., Hancox, G., Polders, E., Brijs, K., Brijs, T., Ross, V., Gruden, C., Šraml, M., Rodošek, V.,
Tollazzi, T., Papadimitriou, E., Lourenco, A., Carreiras, A., & Fortsakis, P. (2020). State of the art on monitoring the
driver state and task demand. Deliverable 2.1 of the Horizon 2020 project i-DREAMS.
13. Manser, M. P., & Hancock, P. A. (2007). The influence of perceptual speed regulation on speed perception, choice, and
control: Tunnel wall characteristics and influences. Accident Analysis & Prevention, 39(1), 69-78.
14. Ranney, T. A., Harbluk, J. L., & Noy, Y. I. (2005). Effects of voice technology on test track driving performance:
Implications for driver distraction. Human factors, 47(2), 439-454.
15. Brooks, J. O., Tyrrell, R. A., & Frank, T. A. (2005). The effects of severe visual challenges on steering performance in
visually healthy young drivers. Optometry and Vision Science, 82(8), 689-697.
16. Greenberg, J., Artz, B., & Cathey, L. (2003). The effect of lateral motion cues during simulated driving. Proceedings of
DSC North America, 8-10.
17. Caird, J. K., Willness, C. R., Steel, P., & Scialfa, C. (2008). A meta-analysis of the effects of cell phones on driver
performance. Accident Analysis & Prevention, 40(4), 1282-1293.
18. Hancock, P. A., Lesch, M., & Simmons, L. (2003). The distraction effects of phone use during a crucial driving maneuver.
Accident Analysis & Prevention, 35(4), 501-514.
19. Farah, H., Polus, A., Bekhor, S., & Toledo, T. (2007). Study of passing gap acceptance behavior using a driving simulator.
Advances in Transportation Studies an International Journal, 9-16.
20. Jia, Y., & Tyler, C. W. (2019). Measurement of saccadic eye movements by electrooculography for simultaneous EEG
recording. Behavior research methods, 51(5), 2139-2151.
21. Liang, Y., Reyes, M. L., & Lee, J. D. (2007). Real-time detection of driver cognitive distraction using support vector
machines. IEEE transactions on intelligent transportation systems, 8(2), 340-350.
22. Mabry, J. E., Glenn, T. L., & Hickman, J. S. (2019). Commercial Motor Vehicle Operator Fatigue Detection Technology
Catalog and Review.
23. Hammoud, R. I., Smith, M. R., Dufour, R., Bakowski, D., & Witt, G. (2008). Driver distraction monitoring and adaptive
safety warning systems (No. 2008-01-2694). SAE Technical Paper.
24. Huang, H., Chen, H., & Lin, S. (2019). MagTrack: Enabling Safe Driving Monitoring with Wearable Magnetics. In
Proceedings of the 17th Annual International Conference on Mobile Systems, Applications, and Services. 326-339.
25. Ji, Q., & Yang, X. (2002). Real-time eye, gaze, and face pose tracking for monitoring driver vigilance. Real-time imaging,
8(5), 357-377.
26. Morimoto, C. H., Koons, D., Amir, A., & Flickner, M. (2000). Pupil detection and tracking using multiple light sources.
Image and vision computing, 18(4), 331-335.
27. Craye, C., Rashwan, A., Kamel, M. S., & Karray, F. (2016). A multi-modal driver fatigue and distraction assessment
system. International Journal of Intelligent Transportation Systems Research, 14(3), 173-194.
28. Jain, J. J., & Busso, C. (2011). Analysis of driver behaviors during common tasks using frontal video camera and CAN-
Bus information. In 2011 IEEE International Conference on Multimedia and Expo. 1-6. IEEE.
29. Li, N., & Busso, C. (2013). Analysis of facial features of drivers under cognitive and visual distractions. In 2013 IEEE
International Conference on Multimedia and Expo (ICME). 1-6. IEEE.
30. Fernández, A., Usamentiaga, R., Carús, J. L., & Casado, R. (2016). Driver distraction using visual-based sensors and
algorithms. Sensors, 16(11), 1805.
31. Edenborough, N., Hammoud, R., Harbach, A., Ingold, A., Kisacanin, B., Malawey, P., & Wil-helm, A. (2005). Driver
state monitor from delphi. In 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2,
1206-1207.
32. Juhola, M., Aalto, H., Joutsijoki, H., & Hirvonen, T. P. (2013). The classification of valid and invalid beats of three-
dimensional nystagmus eye movement signals using machine learning methods. Advances in Artificial Neural Systems.
33. Zimasa, T., Jamson, S., & Henson, B. (2019). The influence of driver’s mood on car following and glance behaviour:
Using cognitive load as an intervention. Transportation research part F: traffic psychology and behaviour, 66, 87-100.
34. Smart eye. (2020, March 21). Eye tracking technology for tomorrows vehicles and research. Retrieved from
https://smarteye.se/.
35. Kumar, D., Dutta, A., Das, A., & Lahiri, U. (2016). Smarteye: developing a novel eye tracking system for quantitative
assessment of oculomotor abnormalities. IEEE Transactions on neural systems and rehabilitation engineering, 24(10),
1051-1059.
36. Optalert. (2020, March 21). Mining - Transport. Retrieved from https://www.optalert.com/industries/mining-transport/.

10
 Michelaraki et al. / RSS2022, Athens, Greece, June 08-10, 2022

37. Corbett, M. A. (2009). Science & Technology Watch: A Drowsiness Detection System for Pilots: Optalert®. Aviation,
space, and environmental medicine, 80(2), 149-149.
38. Cardio Wheel. (2020, March 21). Your Heart Is Smarter Than You Think. Retrieved from https://www.cardio-
id.com/cardiowheel.
39. Texas Instruments. (2020, March 21). Biometric Steering Wheel Reference Design. Retrieved from
http://www.ti.com/lit/ug/tidu479/tidu479.pdf
40. Abu-Faraj, Z. O., Al Chamaa, W., Al Hadchiti, A., Sraj, Y., & Tannous, J. (2018). Design and Development of a Heart-
Attack Detection Steering Wheel. In 2018 11th International Congress on Image and Signal Processing, BioMedical
Engineering and Informatics (CISP-BMEI) 1-6. IEEE.
41. BioRadio by Great Lakes NeuroTechnologies. (2020, March 21). The BioRadio Wireless Physiology Monitor. Retrieved
from https://glneurotech.com/bioradio/.
42. FlexComp (2020, March 21). BioGraph Infiniti Software - Thought Technology. Retrieved from
http://thoughttechnology.com/index.php/flexcomp-system-with-biograph-infiniti-software-t7555m.html
43. Shimmer 3. (2020, March 21). Shimmer 3 GSR unit, EMG3. Retrieved from http://www.shimmersensing.com/.
44. Empatica E4 Wristband. (2020, March 21). Real-time Physiological Signals. Retrieved from
https://www.empatica.com/en-eu/research/e4/.
45. Tobi eye-tracking glasses. (2020, March 21). Tobii Pro Glasses 2. Retrieved from https://www.tobiipro.com/product-
listing/tobii-pro-glasses-2/.
46. Wahlström, J., Skog, I., & Händel, P. (2017). Smartphone-based vehicle telematics: A ten-year anniversary. IEEE
Transactions on Intelligent Transportation Systems, 18(10), 2802-2825.
47. You, C. W., Lane, N. D., Chen, F., Wang, R., Chen, Z., Bao, T. J., & Campbell, A. T. (2013). Car safe app: Alerting
drowsy and distracted drivers using dual cameras on smartphones. In Proceeding of the 11th annual international
conference on Mobile systems, applications and services. 13-26.
48. Yi, D., Su, J., Hu, L., Liu, C., Quddus, M. A., Dianati, M., & Chen, W. H. (2019). Implicit personalization in driving
assistance: State-of-the-art and open issues. IEEE Transactions on Intelligent Vehicles.
49. Azar, S. M., Atigh, M. G., Nickabadi, A., & Alahi, A. (2019). Convolutional Relational Ma-chine for Group Activity
Recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 7892-7901.
50. Formosa, N., Quddus, M., Ison, S., Abdel-Aty, M., & Yuan, J. (2020). Predicting real-time traffic conflicts using deep
learning. Accident Analysis & Prevention, 136, 105429.
51. Valeriano, L. C., Napoletano, P., & Schettini, R. (2018). Recognition of driver distractions using deep learning. In 2018
IEEE 8th International Conference on Consumer Electronics-Berlin (ICCE-Berlin). 1-6.

11