Next Article in Journal
Three-Dimensionally Printed Gypsum Located Within Micro-Encapsulated Phase Change Material: Thermal Conductivity Benefits of Selective Activation Technique
Previous Article in Journal
Offensive Patterns and Performance Analysis in One-Possession Scenarios During the Last Minute and Overtime in the EuroLeague
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Design and Testing of a Wearable System for Monitoring Car Drivers

1
SeaTech, Engineering School, University of Toulon, Av. de l’Université, 83130 La Garde, France
2
Department of Industrial Engineering, Università di Roma Tor Vergata, Via del Politecnico 1, 00133 Rome, Italy
*
Author to whom correspondence should be addressed.
Appl. Sci. 2025, 15(4), 1930; https://doi.org/10.3390/app15041930
Submission received: 16 December 2024 / Revised: 11 February 2025 / Accepted: 12 February 2025 / Published: 13 February 2025
(This article belongs to the Section Electrical, Electronics and Communications Engineering)

Abstract

:
The increasing concern over driver safety has led to the development of various monitoring systems aimed at assessing the physiological state of drivers. This paper presents the design development and testing of a wearable system that monitors key physiological indicators such as heart rate, skin conductance, and movement to evaluate whether the driver’s behavior is aggressive. The system integrates multiple sensors to provide real-time data on the driver’s state. Our objective is to enhance driver safety by detecting aggressive driving behavior through continuous monitoring. This paper covers the conceptual design, performance analysis, and testing layout of the system. Results indicate the system’s effectiveness in detecting changes in physiological states, suggesting its potential for improving road safety and its potential for practical implementation.

1. Introduction

The assessment of a driver’s physiological state is of paramount importance for road safety and the development of advanced monitoring systems. The number of road accidents, even minor ones, is constantly increasing and represents a growing concern for road safety. Indeed, according to the latest data from the World Health Organization (WHO), approximately 1.35 million people die in road accidents each year, while up to 50 million people are injured or disabled. In this context, a thorough understanding of the factors influencing the mental and physical state of drivers is essential for designing effective solutions for detecting and preventing dangerous driving behaviors.
Road accidents represent a major source of injuries, with neck trauma occupying a significant place. Even at low speeds, these incidents can have devastating consequences on the health and well-being of the individuals involved [1]. When a vehicle is rear-ended, the passengers’ necks undergo a series of abrupt movements, putting strain on the delicate structures of this anatomical region [2]. These movements form what is called neck S-shape motion, consisting of three phases, retraction, forward movement, and belt restraint, as modeled in Figure 1. They can lead to injuries ranging from cervical sprains to vertebral fractures, with potentially serious consequences for the mobility and quality of life of the victims.
The monitoring of car driving behavior can help or even prevent car accidents producing serious injuries in car occupants. The modes of those injuries can also give indications on the monitoring of the car driver motion that can be a function of their physiological and psychological status.
The scientific literature has extensively documented the mechanisms of cervical injuries in road accidents. Research has shown that occupant position, seat design, and vehicle speed can all influence the severity of neck injuries [3]. Thus, by analyzing various cases of common accidents, one can identify several injury mechanisms such as hyperextension and hyperflexion of the neck, which occur during rear-end or frontal collisions and result in a sudden backward or forward projection of the head, respectively. Whiplash is a delicate and potentially deadly phenomenon that frequently occurs in this type of accident, although the current understanding of this injury risk remains limited [4]. Even when injuries are not as severe, some accidents can still pose considerable danger. In addition to rear-end and frontal collisions, side collisions can cause excessive rotation of the cervical spine, as often observed with the aspect that can lead to twists of the cervical vertebrae and injuries to surrounding tissues [5]. Considering these additional mechanisms, a deeper understanding of cervical injuries in road accidents can be gained with the aim to improve prevention measures and road safety.
The risks to the neck are not limited to road accidents. Even minor incidents, such as sudden braking, can cause cervical injuries. When the head undergoes rapid forward and backward movement, it can create excessive strain on the soft tissues of the neck, leading to cervical sprains and other similar injuries [6]. Although often considered minor effects, these injuries can have long-term consequences on the health and well-being of affected individuals. Similarly, road accidents can have a significant psychological impact on victims, especially when they suffer from persistent cervical injuries. Likewise, the economic costs associated with cervical injuries are also significant, with substantial annual expenditures on medical treatments and rehabilitation.
In summary, road accidents pose significant health risks to the neck, even at relatively low speeds. Understanding the mechanisms of these injuries is essential for developing effective prevention and treatment strategies. To reduce these accidents and their consequences, innovative solutions have been proposed and studied. On one hand, several studies have focused on analyzing the human body during braking or accidents [7]. On the other hand, many research efforts have been based on designing systems to mitigate these injuries [8]. However, a common starting point for most of these works is the study of the physiological state of the driver, which thus appears to play an essential role in these accidents.
Driver physiology is a complex field that encompasses various aspects of human health, including alertness, fatigue, stress, and emotions. When an individual takes the car wheel, their physiological state can vary depending on numerous factors such as their level of rest, emotional state, and overall health. Understanding these aspects is essential for predicting dangerous behaviors on the road or studying the variation of injury risks based on their physiological state. Recent studies have emphasized the importance of real-time monitoring of the driver’s physiological state, as it directly influences their driving abilities and responsiveness [9].
One can take the example of fatigue, stress, and drowsiness, which significantly reduce the driver’s vigilance, affecting their ability to react quickly to road situations [10]. Indeed, drowsiness while driving is a major problem as it can lead to micro-sleeping, where a driver briefly loses awareness of their surroundings, greatly increasing the risk of serious accidents. Similarly, stressed drivers are more likely to make mistakes such as ignoring a red light or deviating from their car lane, increasing the risk of accidents. Signs of fatigue and stress may be relatively easy to detect, but they can still go unnoticed by the driver themselves. Therefore, it is imperative to design systems capable of real-time monitoring of the driver’s physiological state to prevent potential dangers on the road.
Identifying the physiological signals associated with emotional states is also addressed in the current research on prevention of car accidents. There are a lot of different emotions that can have a real influence on a driver’s behavior. The main emotional states can be recognized in anger, joy, fear, and sadness.
Emotions such as anger, fear, sadness, or joy can impair decision-making and increase the risk of accidents [11]. For example, an angry reaction to another driver can lead to aggressive driving, thereby increasing the risk of collisions. Thus, researching the physiological mechanisms underlying these emotional states of the driver is essential for the development of sophisticated monitoring systems [12]. These systems can help anticipate risky behaviors and prevent accidents on the road.
The monitoring of drivers’ physiological state is a rapidly expanding field, aiming to capture and analyze various aspects of human health and behavior during driving; it is not only limited to detecting fatigue or distraction but also encompasses other emotional and psychological states that can influence road safety.
Technological advancements have led to the development of sophisticated sensors capable of measuring a variety of physiological parameters in real time. These sensors are essential for monitoring the driver’s physiological state, providing critical data that can help improve road safety. To generate an overview of the primary sensors used in this field, we can refer to the following:
  • Cameras and Microphones: Eye-trackers and facial recognition cameras, along with microphones, analyze the driver’s facial and vocal expressions to detect fatigue and emotional states such as anger or frustration. These devices offer valuable insights into how emotional states impact driving behavior [13].
  • Contact Sensors: Devices like heart rate (HR) monitors, heart rate variability (HRV) sensors, respiratory rate (RR) monitors, and skin conductance (SC) sensors collect data that are crucial for assessing driver fatigue and mood changes. These data can indicate the driver’s current state and predict potential issues related to fatigue and stress.
  • Inertial Measurement Units (IMUs): These sensors include accelerometers and gyroscopes that detect driver movements, providing information about movements associated with emotions such as irritation or fear. This information helps in understanding the physical responses related to different emotional states [14].
  • Electroencephalography (EEG): EEG sensors measure brain electrical activity, helping to understand patterns associated with various mental states such as happiness, sadness, anger, and mental load. These data are vital for evaluating the driver’s cognitive and emotional conditions [15].
The abovementioned sensors can be integrated into the driver’s seat or steering wheel, and/or they can be designed as wearable devices like smartwatches, glasses, caps, earpieces, or vests. This integration allows for continuous monitoring of the driver’s physiological state, enabling quick detection and response to changes that may affect their safety and that of other road users.
Once physiological data are collected by sensors, they need to be analyzed to detect patterns, trends, and anomalies [16]. Data analysis methods often include machine learning techniques, signal processing, and pattern recognition. It is with the help of these EEG signals and pattern recognition that one can establish reliable solutions to detect drowsiness [17]. These methods translate raw data into meaningful information about the driver’s health and state of vigilance.
Several studies focus on these key factors which highlight variations in heart rate, respiratory rate, and skin conductance as indicators of stress and anxiety in drivers [11]. Significant advances in fatigue detection have also been made through methods with brainwave analysis and the use of convolutional neural networks [12]. Since fatigue and drowsiness are major factors contributing to road accidents, these analysis techniques are significant for accident prevention.
From the extensive literature on topics referring to the analysis and design of monitoring systems for safe driving, it is worth noting examples of the variety of proposed solutions using sensorization of drivers and cars that are combined with procedures of data elaboration. In [18], the authors present a monitoring system with IMUs and EMG sensor installed on the head and neck to detect drivers’ emotional responses. Reference [19] proposes a detection of abnormal driving by using signals from EMG sensors. In reference [20], a visual detection of drivers is used to evaluate driver distraction by using driver data through neural network algorithms. Driving behavior is formulated in [21] by using signals from distributed sensors that are installed on the driver and on the car. The authors of [22] elaborate on the detection of inattention and drowsiness status of a car driver by using a sensorization of the driver motion and cardiac response.
In the current automotive context, it is significant to assess the driver’s physiological state, particularly focusing on aggressiveness, as it directly impacts driving safety. Aggressive driving behavior can lead to risky maneuvers, increasing accidents, and endangering both the driver and other road users. Aggressiveness can stem from emotions such as stress, joy, and anger, so that it is important and complementary to identify these emotional triggers of aggressive driving. Therefore, developing a means to effectively monitor and interpret signs of aggressiveness in drivers is useful to enhance road safety.
To address this challenge, the objective of this work is to design a portable and discrete system that is capable of properly detecting and analyzing signs of aggressiveness in drivers. Such a system aims to improve driving comfort, safety, and overall road user well-being by providing timely interventions or alerts when aggressive behavior is detected. Using existing research and technological tools, such as IMU sensors and heart rate monitors, gives the possibility to propose a system that can be designed to be compact, unobtrusive, and easily deployable without the need for permanent installation in vehicles. Integrating multiple sensors, including IMU sensors for motion detection and heart rate monitors for physiological indicators, will enable a comprehensive assessment of the driver’s aggressiveness levels. Furthermore, the development of real-time data processing algorithms is essential to promptly identify and respond to instances of aggressiveness, potentially through fatigue warnings or adjustments to driver assistance systems.
With a clear focus on detecting and addressing aggressive behavior in drivers, this paper presents a monitoring system that aims to contribute to the enhancement of road safety and the optimization of driving experiences for all road users.
The aim of the work can be summarized in the discussion of problems and solutions faced when designing a proper sensor-based system to detect the physiological state of a car driver and its impact on their proper driving behavior by using a minimal set of sensors in a comfortable wearable system. The paper refers to the area of applied biosciences and bioengineering and it is organized as follows: Section 1 presents the subjects of interest in the paper, looking at the main problems and current solutions; Section 2 presents the motivations for the work and the proposed solution of a motoring system of the state of a car driver; Section 3 discusses test results to validate and characterize the proposed monitoring system; and Section 4 gives comments on the challenges and potential practical use of the proposed monitoring system for car drivers.

2. Materials and Methods

2.1. Motivations and Requirements

The primary motivation of this work is to enhance road safety by monitoring and analyzing driver behavior, particularly focusing on aggressive driving. Aggressive driving significantly increases the risk of road accidents, endangering both the driver and other road users. Emotions such as stress, anger, and joy can trigger aggressive driving behaviors, making it essential to monitor these physiological and psychological states in real time.
A monitoring system must meet several significant requirements including the following:
  • Portability and Discreteness: The device should be compact, unobtrusive, and easily deployable without requiring permanent installation in the vehicle frames.
  • Data Accuracy and Reliability: It should accurately collect and analyze physiological data to reliably detect aggressive driving behaviors.
  • Safety and Compliance: The device must adhere to safety protocols and regulatory standards for wearable technology and data privacy.
  • Real-time Processing: The system should process data in real time to provide timely alerts or interventions.
  • Cost-effectiveness: The overall system should be cost-effective to make it accessible for widespread adoption.

2.2. Conceptual Design

A monitoring system is proposed in this paper that consists of three wearable device supports, namely a wristband, a headband, and a torso patch, as indicated in Figure 2. The wristband on the wrist serves as the central hub, containing several sensors, while the headband and torso patch house mainly IMU sensors. The used IMU sensors are IMU MPU6050 (2.5 × 3 × 0.83 mm, +16 g range) [23], integrated in Arduino 33 IoT [24], the heart rate sensor Analog MAX30102 (5.6 × 3.3 × 1.55, 1000 pulse/s range) [25], and the galvanic skin Grove GSR sensor (20 × 20 mm and 2000 mS range) [26]. The wristband houses a heart rate sensor and a skin response GSR sensor that can be used to monitor physiological signals indicating stress and aggressiveness. The wristband is provided with an IMU to monitor the arm motion at the level of the wrist. The torso unit uses an IMU to monitor the motion of the torso; likewise, the head unit is equipped with another IMU. Each unit is provided with an Arduino 33 IoT (45 × 18 × 2 mm) [18] for data acquisition and handling.
The selected sensors communicate wirelessly via Bluetooth to transmit data to a central processing unit, according to the scheme in Figure 3. The system is powered by rechargeable batteries, ensuring uninterrupted operation during driving. Each unit can be considered independent as they are all equipped with a proper Nano Arduino board and power source battery. However, the data flow can be shared among the three Nano Arduino boards via Bluetooth in order to be used for data elaboration, storage, and alerts.
The data acquisition focuses on evaluating the system’s ability to accurately capture and process physiological data, particularly in detecting aggressive driving behaviors using data coming from the wristband, torso, and headband units. Several tests were carried out to ensure the sensors’ accuracy, synchronization, and reliability. In particular, the test experiences gave the following results.
The heart rate sensor MAX30102 had some difficulties in accurately capturing heart rate variability (HRV) as a significant indicator of stress and aggression. In addition, its difficult installation reduced its reliability, meaning it was not completely successful.
The skin conductance sensor (Grove GSR) also had some difficulties in accurately measuring skin conductance data as an indicator of the driver’s emotional state. In addition, its difficult installation reduced its reliability, meaning it was not completely successful.
The IMU sensors in the wristband, torso box, and headband were successful in properly detecting the movements of the body parts where they were installed. The data showed consistent and reliable detection of acceleration and angular velocity, as significant measures for identifying sudden and aggressive driving maneuvers.

2.3. Testing Layout and Modes

The testing layout was designed to simulate real-world driving conditions and to capture data related to aggressive driving behaviors. The setup included strategic placement of sensors on the driver’s body and the use of different test modes to simulate various driving scenarios.
The sensors were placed to capture relevant physiological and movement data effectively in the following locations:
  • Wristband: it is installed on the wrist to monitor heart rate, skin conductance, and the wrist movement for sudden steering wheel maneuvers.
  • Torso Box: it is attached to the torso to capture upper body movements.
  • Headband: it is worn on the head to measure head movements.
In particular, the IMU data from the torso box and the headband will also provide information on neck acceleration, indicating aggressive driving due to large movements of the torso and head.
The location of the sensor units is indicated in Figure 4 with a solution ensuring a wearing solution without excessive discomfort for a car driver.
The system was tested in various modes to ensure data collection for the following conditions:
  • Normal Driving: Tests were carried out under standard driving conditions to establish a reference for physiological and movement parameters using all the device sensors.
  • Aggressive Driving: A driver performed aggressive maneuvers such as rapid accelerations, hard braking, and sharp turns. This mode aimed to capture the physiological and movement responses associated with aggressive driving.
To proceed with the various test modes, the following experimental procedure was implemented:
  • Baseline Measurement: Initial measurements were taken during a period of calm driving to establish baseline physiological and movement data.
  • Simulated Driving Modes: A driver was instructed to perform aggressive driving maneuvers while the system recorded heart rate, GSR, and IMU data.
  • Data Analysis: The collected data were analyzed to determine if any parameter exceeded its threshold for calm driving by using a criterion value for the driver’s level of aggressiveness.
A criterion for the driver’s level of aggressiveness can be proposed to quantify driving aggressiveness using the parameters acquired by the sensors of the proposed monitoring device as follows:
A D C = W n   a n + W w   a w + W h   r h + W s   c s
where Wn is the weighing coefficient for the value of the neck acceleration an; Ww is the weighing coefficient for the value of the wrist acceleration aw ; Wh is the weighing coefficient for the value of the heart rate rh; and Ws is the weighing coefficient for the value of the skin conductance cs. Table 1 summarizes the components of the proposed aggressive driving criterion (ADC), including each parameter’s threshold value, weighting coefficients Wi, and the rationale for its weighting. As reported in Table 1, the index ADC combines motion data with heart rate and changes in skin conductance due to sweating.
While, from a medical perspective, heart rate and sweating can be accurate markers of aggressive behavior, the quality of the relative sensor acquisitions is strongly influenced by sensor placement, since a wrong installation can result in non-usable data or false positives. Conversely, while wrist and neck motion could be caused by different behavior, the relative sensing systems, based on IMUs, are easy to install and reliable. Because of this, the aggressive driving criterion weighs IMU measurements more than heart rate and skin conductance. Threshold values for each sensor, dividing non-aggressive driving from aggressive driving, have been selected by integrating the existing literature, as shown in the Introduction Section, with observations from repeated experiments. Overall, the proposed ADC index formulates a comprehensive assessment of the driver’s aggressive behavior by integrating multiple physiological and movement indicators into a single parameter, which can be used for both monitoring and intervention through sound or haptic warnings. The criterion is used by indicating an aggressive driving behavior when any one of the parameters overpasses its threshold, whereas the criterion value gives an overall estimation.

3. Results

Several tests were carried out and repeated three times with successful results, as in the reported illustrative example in Figure 5, Figure 6 and Figure 7.

3.1. Experimental Setup

The device sensors were installed on a volunteer car driver, and data were collected according to the scheme in Figure 3. The driver followed a predefined sequence of movements, alternating motion between calm and aggressive driving, referring to accelerating and braking in a straight line, turning left, turning right, and then back to straight-line acceleration and braking.
In the first mode of test, the volunteer’s left forearm was sensorized by the wristband that was equipped with a heart rate sensor (MAX30102) and a skin conductance sensor (Grove GSR) as in the scheme of Figure 4. The Arduino Nano was connected to an added TCA9548A multiplexer and MPU6050 IMU sensor [23].
In the second mode of test, the volunteer was sensorized to a setup mimicking a jacket with two IMU sensors to measure head and torso accelerations as in the scheme of Figure 4. An Arduino Nano, TCA9548A multiplexer, and two MPU6050 IMU sensors were used to capture neck acceleration.
The setup of both test modes is illustrated in Figure 5. In both test modes the volunteer car driver performed the planned movements while acquiring sensor data.

3.2. Test Results

The acquired data from both the wristband and the upper body device in both test modes in Figure 5 were analyzed to evaluate calm and aggressive driving behavior.
Referring to the calm driving in the first test mode, the results are shown in Figure 6, indicating that none of the threshold values of the aggressive driving criterion in Table 1 were exceeded, resulting in an evaluation of the criterion for non-aggressive driving. Sensor data from the MAX30102 heart rate sensor, Grove GSR, and IMUs were recorded as reported in Figure 6a with a regular time without exceeding thresholds, confirming the effectiveness of the system in detecting calm driving. Figure 6b shows the acceleration magnitudes of the torso and neck with time evolution and values well below the threshold for aggressive driving according to the aggressive driving criterion in Table 1. The torso acceleration is quite constant, indicating that the driver did not move much during the test, while the neck acceleration, recorded by the head IMU sensor, shows motions that can be related to turning the head when looking around during driving or to following the motion of steering when turning the car on the road.
Referring to the aggressive driving in the second test mode, the results are shown in Figure 7, indicating that several acquired data points exceed the threshold values. The results in Figure 7a for heart rate (in pulse/s) and skin conductance (in mS) do not show significant variation and have low values, possibly due to failure in the sensors’ acquisition or driver acclimatization to aggressive driving. However, both neck and wrist acceleration in Figure 7b surpassed their thresholds, as the red circle highlights, meeting the criterion evaluation for aggressive driving.
It is worth noting the similarity of the behavior in the experienced test mode in terms of the shape of the time evolution of the acquired data; however, the different car driving attitudes result mainly in a difference in the numerical values of the acquired data, as highlighted in terms of the acceleration of the neck and head of the car driver.
In summary, while the IMU data effectively distinguish between calm and aggressive driving, data from the heart rate and skin conductance sensors do not show such a significant detection as expected, very likely due to physiological limitations, such as the need for more long-time monitoring and better sensor capabilities.

4. Discussion

The experience carried out with the prototype of the proposed monitoring device, as reported with the results of the discussed tests, demonstrates potential and limitations in the efficiency and operation of the designed solution. In particular, the potential of the device can be considered in terms of easy wearing and comfort without impediments for car drivers, as shown in the example in Figure 5. The limited size of the units, each contained in a box of approximately 5.5 × 5.0 × 2.0 cm, also ensures easy portability. Even the relatively low cost, which can be estimated overall at less than EUR 100 for the prototype, much lower than that of existing systems, allows us to foresee the potential success of practical implementation for both vehicle and truck drivers. Limits for practical use can be identified in its structural and functional complexity due to its composition of three independent units; however, these can be easily maintained and managed. The potential of the proposed device is also related to the wide monitoring area with the parameters of motion and physiological reactivity when compared to existing systems which focus mainly on states of drowsiness of vehicle drivers. Data management with direct acquisition of signals and evaluation of the state of aggressiveness by comparison with threshold values makes the ADC calculated using Equation (1) and displayed in Table 1 easy to use with immediate effect. This also presents the possibility of further developments for sound or visual alarm systems that can also be integrated into the same units of the device. This informatic aspect characterizes the functional efficiency of the device, which was designed for use in a system to monitor the aggressive driving of vehicles, due to various reasons, but the physiological monitoring is achievable with few parameters. The parameters used were found to be useful and very efficient in monitoring the driver movement resulting from aggressive conditions, while those more dedicated to the physiological reactivity of the heartbeat and skin reaction did not give the desired results due to evident indications of variations in physiological state. Although the heart rate and skin conductance sensors were found to be inefficient in the proposed device, this evidence can still be considered a significant result of monitoring research and design in the sense that it indicates the need for better sensors of that type or to examine alternatives. The exclusion of heart rate and skin reactivity monitoring in existing devices is not documented by any other studies, and therefore, the result reported in this work can be considered interesting for this purpose.
Implementing a monitoring system comprising separate devices, such as the proposed wristband, torso box, and headband, presents several practical challenges despite their individual autonomy. These challenges include the following:
  • Device Management and Maintenance: Each device requires periodic recharging, which can be cumbersome for users, particularly during long driving sessions. Managing multiple devices introduces logistical complexities, as users must ensure that all components are charged and functioning before use.
  • Comfort and Usability: Although the three device units are designed to be lightweight and wearable for long periods, they could lead to discomfort while driving. The physical presence of the wristband, torso box, and headband may cause distraction or irritation, potentially impacting the driver’s experience and driving performance.
  • Calibration and Sensor Reliability: Calibration is a critical issue to ensure accurate measurements for significant data elaboration. However, discrepancies and malfunctions can occur, necessitating frequent recalibration of sensors. This issue can affect the reliability of the data and the system’s overall performance.
  • Cost Considerations: As a prototype, the current system involves significant costs, primarily due to the design activity, used sensors, built additional parts, and testing setup and lab experiences. The expense may be a barrier to widespread adoption and implementation, highlighting the need for cost-effective alternatives or improvements in manufacturing processes.
Despite the abovementioned challenges, the proposed monitoring system and the data evaluation through the aggressive driving criterion can be considered a contribution to improve safe driving. Using low-cost, comfortable, wearable sensors on the driver, fairly simple data collection can be carried out to analyze aggressive driving when compared with existing solutions with complex structures, limited tasks referring mainly to drowsiness detection, or heavy computation for data analysis of driving behavior.
In summary, while the prototype demonstrates significant potential for real-time detection of aggressive driving by using different device units, practical implementation will address issues related to device management, user comfort, calibration, and cost. Future developments should focus on enhancing ease of use, improving sensor accuracy, and reducing costs to make the system more viable for practical applications.
Additionally, while the experiments demonstrate the feasibility of the proposed aggressive driving criterion (ADC), the numerical value of weights and the threshold can be improved with an extensive experimental campaign, since current values have been selected according to a limited number of tests and subjects. Future testing should include different driving situations, analyzing how these thresholds vary according to road conditions, the average speed of driving, the environment (e.g., city, suburbs, countryside, highways), and travel duration (e.g., short daily commute, long-distance driving, driving with frequent stops, etc.).

5. Conclusions

This paper presents a wearable monitoring system for car drivers that is aimed at enhancing road safety by detecting aggressive driving behaviors. The proposed system integrates inertial measurement units (IMUs), heart rate monitors, and skin conductance sensors to continuously collect and analyze physiological and movement data of a driver during car driving. The aggressive driving criterion (ADC) is formulated to evaluate the aggressiveness status while driving by using the designed monitoring system. The test results show the system’s effectiveness in distinguishing between calm and aggressive driving by mainly using data from IMU sensors to capture neck and wrist accelerations. The heart rate and skin conductance from sensors on a wristband do not exhibit significant variation between driving conditions, suggesting that it would be more convenient to implement better sensors with better calibration or use sensors with alternative physiological parameters. Future improvements should focus on improving sensor accuracy and addressing the identified weaknesses to fully monitor the drivers’ conditions.

Author Contributions

Conceptualization, M.C.; Methodology, J.M., M.C. and M.R.; Software, J.M.; Validation, J.M.; Formal analysis, J.M. and M.R.; Investigation, M.C.; Resources, M.C.; Data curation, J.M.; Writing—original draft, J.M. and M.C.; Writing—review & editing, M.C. and M.R.; Visualization, J.M.; Supervision, M.C.; Project administration, M.C.; Funding acquisition, M.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in the study are included in the article; further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Ono, K.; Kanno, M. Influences of the Physical Parameters on the Risk to Neck Injuries in Low Impact Speed Rear-end Collisions. Accid. Anal. Prev. 1996, 28, 493–499. [Google Scholar] [CrossRef]
  2. Garrosa, M.; Ceccarelli, M.; Diaz, V.; Russo, M. Experimental Validation of a Driver Monitoring System. Machines 2023, 11, 1060. [Google Scholar] [CrossRef]
  3. Zuby, D.; Lund, A. Preventing Minor Neck Injuries in Rear Crashes—Forty Years of Progress. J. Occup. Environ. Med. 2010, 52, 428–433. [Google Scholar] [CrossRef]
  4. Tencer, A.F.; Huber, P.; Mirza, S.K. A Comparison of Biomechanical Mechanisms of Whiplash Injury from Rear Impacts. Annual proceedings. Assoc. Adv. Automot. Med. 2003, 47, 383–398. Available online: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3217546/ (accessed on 9 January 2025). [PubMed]
  5. Kang, Y.-S.; Stammen, J.; Bendig, A.; Agnew, A.; Hagedorn, A.; Thomas, C.; Ramachandra, R.; Kwon, H.-J.; Moorhouse, K.; Bolte, J. Effects of Seatback Recline and Belt Restraint Type on PMHS Responses and Injuries in Rear-Facing Frontal Impacts. SAE Int. J. Transp. Saf. 2022, 10, 235–290. [Google Scholar] [CrossRef]
  6. Garrosa, M.; Ceccarelli, M.; Diaz, V. Problems and Requirements in Impact Analysis from Vehicle Accidents. Adv. Ital. Mech. Sci. 2022, 122, 346–354. [Google Scholar] [CrossRef]
  7. Laribi, M.-A.; Ceccarelli, M.; Nelson Carl, A.; Zeghloul, S. Torso Motion Monitoring with an IMU Set-Up. In New Advances in Mechanisms, Transmissions and Applications: Proceedings of the Sixth MeTrApp Conference 2023; Springer: Berlin/Heidelberg, Germany, 2023; pp. 65–73. [Google Scholar] [CrossRef]
  8. Lundell, B.; Jakobsson, L.; Alfredsson, B.; Jernström, C.; Isaksson-Hellman, I. Guidelines for and the Design of a Car Seat Concept for Improved Protection Against Neck Injuries in Rear End Car Impacts; SAE Technical Paper 980301; Springer: Cham, Switzerland, 1998; pp. 653–663. [Google Scholar] [CrossRef]
  9. Davoli, L.; Mattioli, V.; Gambetta, S.; Belli, L.; Carnevali, L.; Martaló, M.; Sgoifo, A.; Raheli, R.; Ferrari, G. Non-Invasive Psycho- Physiological Driver Monitoring through IoT-Oriented Systems. The Internet of Medical Things: Enabling technologies and emerging applications. Healthc. Technol. 2021, 19–33. [Google Scholar] [CrossRef]
  10. Huang, J.; Huang, X.; Peng, Y.; Hu, L. Driver state recognition with physiological signals Based on deep feature fusion and feature selection techniques. Biomed. Signal Process. Control 2024, 93, 106–204. [Google Scholar] [CrossRef]
  11. Begum, S. Intelligent driver monitoring systems based on physiological sensor signals: A review. In Proceedings of the 16th International IEEE Conference on Intelligent Transportation Systems (ITSC 2013), The Hague, The Netherlands, 6–9 October 2013; pp. 282–289. [Google Scholar] [CrossRef]
  12. Sriranga, A.-K.; Lu, Q.; Birrell, S. A Systematic Review of In-Vehicle Physiological Indices and Sensor Technology for Driver Mental Workload Monitoring. Sensors 2023, 23, 2214. [Google Scholar] [CrossRef]
  13. Kim, D.; Park, H.; Kim, T.; Kim, W.; Paik, J. Real-time driver monitoring system with facial landmark-based eye closure detection and head pose recognition. Sci. Rep. 2023, 13, 18264. [Google Scholar] [CrossRef]
  14. Garrosa, M.; Russo, M.; Ceccarelli, M.; Cafolla, D. Lab Experiences for a Driver Monitoring System. In New Trends in Medical and Service Robotics; Springer: Warsaw, Poland, 2023; pp. 109–116. [Google Scholar] [CrossRef]
  15. Ying, M.; Shao, X.; Zhu, J.; Zhao, Q.; Li, X.; Hu, B. EDT: An EEG-based attention model for feature learning and depression recognition. Biomed. Signal Process. Control 2024, 93, 106–182. [Google Scholar] [CrossRef]
  16. Jeong, M.; Kim, D.; Park, S.; Paik, J. Drowsy status monitoring system: Based on face feature analysis. In Proceedings of the 2022 International Conference on Electronics, Information, and Communication (ICEIC) 1–4 (IEEE, 2022), Jeju, Republic of Korea, 6–9 February 2022; pp. 1–4. [Google Scholar] [CrossRef]
  17. Zhang, K.; Wu, D.; Liu, Q.; Dong, F.; Liu, J.; Jiang, L.; Yuan, Y. Algorithm for drowsiness detection based on hybrid brain network parameter optimization. Biomed. Signal Process. Control 2024, 94, 106–204. [Google Scholar] [CrossRef]
  18. Lee, B.G.; Chong, T.W.; Lee, B.L.; Park, H.J.; Kim, Y.N.; Kim, B. Wearable Mobile-Based Emotional Response-Monitoring System for Drivers. IEEE Trans. Hum.-Mach. Syst. 2017, 47, 636–649. [Google Scholar] [CrossRef]
  19. Fan, Y.; Gu, F.; Wang, J.; Wang, J.; Lu, K.; Niu, J. SafeDriving: An Effective Abnormal Driving Behavior Detection System Based on EMG Signals. IEEE Internet Things J. 2022, 9, 12338–12350. [Google Scholar] [CrossRef]
  20. Khan, T.; Choi, G.; Lee, S. EFFNet-CA: An Efficient Driver Distraction Detection Based on Multiscale Features Extractions and Channel Attention Mechanism. Sensors 2023, 23, 3835. [Google Scholar] [CrossRef]
  21. Bonfati, L.V.; Mendes Junior, J.J.A.; Siqueira, H.V.; Stevan, S.L., Jr. Correlation Analysis of In-Vehicle Sensors Data and Driver Signals in Identifying Driving and Driver Behaviors. Sensors 2023, 23, 263. [Google Scholar] [CrossRef]
  22. Akiduki, T.; Nagasawa, J.; Zhang, Z.; Omae, Y.; Arakawa, T.; Takahashi, H. Inattentive Driving Detection Using Body-Worn Sensors: Feasibility Study. Sensors 2022, 22, 352. [Google Scholar] [CrossRef]
  23. InvenSence, IMU Datasheet. Available online: https://invensense.tdk.com/wp-content/uploads/2015/02/MPU-6000-Datasheet1.pdf (accessed on 9 January 2025).
  24. Arduino, Arduino 33 IoT Datasheet. Available online: https://content.arduino.cc/assets/st_imu_lsm6ds3_datasheet.pdf (accessed on 9 January 2025).
  25. Analog, MAX30102 Datasheet. Available online: https://www.analog.com/media/en/technical-documentation/data-sheets/MAX30102.pdf (accessed on 9 January 2025).
  26. Seeed Studion, Grove GSR Sensor Datasheet. Available online: https://wiki.seeedstudio.com/Grove-GSR_Sensor/ (accessed on 9 January 2025).
Figure 1. Phases of neck S-shape motion in a rear impact [2]: (a) retraction; (b) forward movement; (c) belt restraint.
Figure 1. Phases of neck S-shape motion in a rear impact [2]: (a) retraction; (b) forward movement; (c) belt restraint.
Applsci 15 01930 g001
Figure 2. Location of the main sensors in the proposed monitoring system on the body parts of a car driver.
Figure 2. Location of the main sensors in the proposed monitoring system on the body parts of a car driver.
Applsci 15 01930 g002
Figure 3. Conceptual design of the proposed solution with three wearable units hosting sensors and other equipment communicating data via Bluetooth.
Figure 3. Conceptual design of the proposed solution with three wearable units hosting sensors and other equipment communicating data via Bluetooth.
Applsci 15 01930 g003
Figure 4. Location of the sensor units on a car driver: wristband with monitor heart rate sensor, skin conductance sensor, and IMU sensor; torso box with IMU sensor; and headband with IMU sensor.
Figure 4. Location of the sensor units on a car driver: wristband with monitor heart rate sensor, skin conductance sensor, and IMU sensor; torso box with IMU sensor; and headband with IMU sensor.
Applsci 15 01930 g004
Figure 5. Layout of test setup with a volunteer as in Figure 3: (a) in the first mode; (b) in the second mode.
Figure 5. Layout of test setup with a volunteer as in Figure 3: (a) in the first mode; (b) in the second mode.
Applsci 15 01930 g005
Figure 6. Acquired data during first test mode for calm driving: (a) heart rate in BPM and skin conductance in GSR from wristband device; (b) torso and neck acceleration from torso box and headband.
Figure 6. Acquired data during first test mode for calm driving: (a) heart rate in BPM and skin conductance in GSR from wristband device; (b) torso and neck acceleration from torso box and headband.
Applsci 15 01930 g006
Figure 7. Acquired data during second test mode for aggressive driving: (a) heart rate in BPM and skin conductance in GSR from wristband device: (b) torso and neck acceleration from torso box and headband.
Figure 7. Acquired data during second test mode for aggressive driving: (a) heart rate in BPM and skin conductance in GSR from wristband device: (b) torso and neck acceleration from torso box and headband.
Applsci 15 01930 g007
Table 1. Parameter thresholds and weighting coefficients for the proposed aggressive driving criterion (ADC).
Table 1. Parameter thresholds and weighting coefficients for the proposed aggressive driving criterion (ADC).
ParameterThresholdWeighting CoefficientReason
Neck acceleration4.0 m/s20.3High reliability and significant indicator of head and torso movements
Wrist acceleration2.0 m/s20.3High reliability and significant indicator for detecting aggressive steering maneuvers
Heart rate20 BPM0.2Medium reliability due to difficulties in installation and data accuracy and significant in emotional state
Skin conductance1000 μS0.2Medium reliability due to difficulties in installation and data accuracy and significant in stress state
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Moinard, J.; Ceccarelli, M.; Russo, M. Design and Testing of a Wearable System for Monitoring Car Drivers. Appl. Sci. 2025, 15, 1930. https://doi.org/10.3390/app15041930

AMA Style

Moinard J, Ceccarelli M, Russo M. Design and Testing of a Wearable System for Monitoring Car Drivers. Applied Sciences. 2025; 15(4):1930. https://doi.org/10.3390/app15041930

Chicago/Turabian Style

Moinard, Julien, Marco Ceccarelli, and Matteo Russo. 2025. "Design and Testing of a Wearable System for Monitoring Car Drivers" Applied Sciences 15, no. 4: 1930. https://doi.org/10.3390/app15041930

APA Style

Moinard, J., Ceccarelli, M., & Russo, M. (2025). Design and Testing of a Wearable System for Monitoring Car Drivers. Applied Sciences, 15(4), 1930. https://doi.org/10.3390/app15041930

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop