In-Vehicle Driver Health Monitoring and Alarming System

01 · Project Information

• Project Duration: April 2022 - May 2024

• Role: Team Leader

• Advisor: Prof. Yi Zhang, Southwest Jiaotong University

• Project Type: National Research Project

02 · Project Overview

This project aimed to develop an advanced in-vehicle driver health monitoring and alarming system capable of identifying early signs of driver fatigue and hazardous environmental conditions.

The system combined real-time environmental monitoring with physiological data acquisition to evaluate driver well-being and maintain suitable in-vehicle conditions. When physiological or environmental parameters deviated from their expected ranges, the system identified the potential risk and activated an appropriate intervention.

The key innovations included:

• Real-time monitoring of in-vehicle temperature and carbon dioxide levels

• Non-intrusive ECG acquisition using fabric electrodes

• Fatigue analysis based on heart rate and heart rate variability

• Multi-sensor fusion to reduce false alarms

• Multi-modal alerts for driver intervention

• Wireless communication and edge processing for real-time operation

03 · Project Background

Driver fatigue is affected not only by sleep and driving duration but also by environmental conditions and physiological changes.

Excessive carbon dioxide levels and unsuitable temperatures can contribute to drowsiness, reduced attention, and slower reaction times. Fatigue may also be reflected through changes in heart rate and heart rate variability.

Traditional driver-monitoring systems often depend on cameras, wearable devices, or individual sensors. Their effectiveness may be limited by lighting conditions, wearing comfort, or the reliability of a single data source.

This project combined environmental parameters with non-intrusive ECG signals to provide a more comprehensive evaluation of driver condition.

04 · Project Objectives

The primary objectives were to:

1. Develop a real-time system for monitoring in-vehicle environmental and physiological conditions

2. Achieve high-precision data acquisition through non-intrusive sensing

3. Detect driver fatigue and hazardous environmental conditions

4. Implement a real-time alerting mechanism

5. Optimize the system for reliable monitoring and intervention

6. Develop a scalable architecture for future applications

05 · System Architecture

The system consisted of the following modules:

• In-vehicle environmental monitoring

• Non-contact ECG acquisition

• Wireless data transmission

• Physiological and environmental data processing

• Fatigue and anomaly detection

• Multi-sensor data fusion

• Multi-modal driver alerts

• Ventilation and air-conditioning control

Sensors continuously collected environmental and physiological data and transmitted the readings to an onboard processing unit. The processing system analyzed the parameters and triggered an alert or environmental adjustment when it detected fatigue or hazardous conditions.

06 · Technical Implementation

1. Environmental Monitoring System

A real-time monitoring system was implemented to maintain a safe and comfortable in-vehicle environment.

The primary monitored parameters included:

• Temperature: Maintained within 5% of optimal conditions to reduce thermal discomfort

• Carbon dioxide levels: Monitored to prevent excessive buildup and reduce the risk of drowsiness

The environmental monitoring module provided:

• Real-time temperature acquisition

• Real-time carbon dioxide monitoring

• Wireless transmission of sensor readings

• Communication with the onboard processing unit

• Evaluation of environmental conditions

• Automatic ventilation and air-conditioning adjustment

• Transmission of abnormal conditions to the alerting module

The regulation mechanism enabled the system to respond to environmental risks instead of only reporting them.

2. Non-Intrusive ECG Acquisition

A non-contact ECG acquisition system based on fabric electrodes was developed to continuously monitor cardiac activity and detect early fatigue-related changes.

The fabric electrodes were integrated into the driver's seat, reducing the need for additional wearable equipment and minimizing interference with normal driving behavior.

The module provided:

• Fabric-electrode integration within the driver's seat

• Comfortable and non-intrusive signal collection

• Continuous ECG monitoring

• Heart-rate extraction

• Heart-rate variability analysis

• Detection of subtle fatigue-related physiological changes

The system achieved 97% ECG signal fidelity, providing a reliable physiological foundation for fatigue detection.

3. ECG Signal Processing

Vehicle vibration, driver movement, and electromagnetic interference can introduce noise into non-intrusively acquired ECG signals.

Adaptive filtering was applied to improve signal quality.

The processing workflow included:

1. Acquiring the raw ECG signal

2. Checking signal integrity

3. Removing clearly abnormal data

4. Applying adaptive filtering

5. Extracting heart-rate information

6. Analyzing heart-rate variability

7. Delivering the physiological indicators to the fatigue-detection module

The signal-processing module improved the usability of ECG data and reduced the influence of motion artifacts and environmental noise.

4. Fatigue Detection

The system analyzed subtle changes in heart rate and heart-rate variability to identify potential driver fatigue or drowsiness.

The fatigue assessment incorporated:

• Real-time heart rate

• Heart-rate variability

• ECG trends

• In-vehicle temperature

• Carbon dioxide concentration

• Baseline driver health parameters

• Environmental parameter trends

Combining physiological and environmental information enabled the system to evaluate driver condition from multiple perspectives.

5. Multi-Sensor Data Fusion

Depending on a single sensor could cause false alerts when temporary fluctuations occurred.

The project therefore adopted a multi-sensor fusion approach that combined driver physiological data with in-vehicle environmental readings.

The system evaluated relationships such as:

• Whether heart-rate changes occurred alongside increasing temperature

• Whether abnormal HRV was accompanied by rising carbon dioxide levels

• Whether an individual sensor anomaly persisted

• Whether multiple indicators exceeded their thresholds simultaneously

The fusion of multiple data sources improved the reliability of fatigue and environmental risk detection.

6. Dynamic Threshold Adjustment

Temperature, carbon dioxide concentration, and physiological parameters naturally change over time and across driving conditions.

To reduce false alarms caused by fixed thresholds, the system introduced dynamic threshold adjustment based on real-time data and baseline parameters.

The threshold mechanism considered:

• Baseline physiological conditions

• Normal environmental fluctuations

• The degree of deviation from optimal conditions

• The duration of an abnormal condition

• Simultaneous changes across multiple indicators

The intervention thresholds were tuned to trigger alerts primarily under genuine fatigue or hazardous conditions.

7. Real-Time Alerting Mechanism

A multi-modal fatigue and environmental alerting mechanism was integrated into the system.

When the system detected potential fatigue or hazardous environmental parameters, it could notify the driver through:

• Audio alerts

• Seat vibration

• Dashboard notifications

• Environmental warning messages

• Ventilation and air-conditioning adjustment

The use of multiple alert channels reduced the likelihood that the driver would overlook a single warning.

8. Wireless Communication and Edge Processing

The environmental and physiological sensing modules transmitted real-time data wirelessly to the onboard processing unit.

Edge processing was used to perform part of the data preprocessing and state evaluation locally, reducing communication latency and improving processing efficiency.

The design included:

• Real-time sensor data transmission

• Wireless communication status monitoring

• Local signal preprocessing

• Reduced transmitted data volume

• Lower processing latency

• Faster response to abnormal conditions

This architecture improved the real-time performance and independent operating capability of the system.

07 · System Workflow

The system followed the process below:

1. Start the in-vehicle monitoring system

2. Initialize the environmental and physiological sensors

3. Measure temperature and carbon dioxide levels

4. Acquire ECG signals through the fabric electrodes

5. Validate and preprocess the raw data

6. Extract heart rate and heart-rate variability

7. Fuse environmental and physiological information

8. Compare the readings with baseline values and dynamic thresholds

9. Determine whether fatigue or an environmental hazard exists

10. Trigger audio, vibration, or dashboard alerts

11. Activate ventilation or air-conditioning adjustments when necessary

12. Continue monitoring and updating the data

08 · Challenges and Solutions

09 · My Responsibilities

As the team leader, I was responsible for:

• Defining project objectives and the overall research plan

• Dividing the system into environmental, ECG, and alerting modules

• Coordinating team tasks and project progress

• Participating in the overall system architecture

• Supporting the development of the environmental monitoring module

• Participating in the fabric-electrode ECG acquisition design

• Organizing signal-acquisition and quality tests

• Supporting fatigue detection and multi-sensor fusion research

• Managing module integration

• Organizing system testing, result analysis, and project presentations

• Preparing research materials and milestone deliverables

10 · Key Contributions and Outcomes

• Designed a driver health-monitoring system combining environmental and physiological data

• Implemented real-time temperature and carbon dioxide monitoring

• Maintained environmental parameters within 5% of optimal conditions

• Developed fabric electrodes integrated into the driver's seat

• Implemented non-intrusive ECG acquisition

• Achieved 97% ECG signal fidelity

• Analyzed driver fatigue through heart rate and heart-rate variability

• Applied adaptive filtering to improve ECG signal quality

• Reduced false-alert risks through multi-sensor fusion

• Implemented audio, seat-vibration, and dashboard alerts

• Designed a scalable architecture for commercial vehicles and fleets

11 · Project Impact

The project combined driver physiological monitoring, environmental sensing, and real-time alerts within a unified system, providing multiple data sources for the early identification of driver fatigue.

Compared with monitoring methods that depend exclusively on cameras or wearable devices, seat-integrated fabric electrodes enabled continuous cardiac monitoring with reduced driver interference. Environmental sensing also accounted for the influence of temperature and carbon dioxide on driver alertness.

The system architecture has the potential to be integrated into commercial vehicles and transportation fleets to improve driver safety management and operational reliability.

12 · Conclusion

This project developed an in-vehicle driver health monitoring and alarming system that integrated non-intrusive ECG acquisition, environmental sensing, wireless communication, multi-sensor fusion, and real-time alerts.

Through the project, I strengthened my knowledge of physiological signal acquisition, ECG processing, wireless sensing, environmental monitoring, and embedded-system integration. I also developed practical experience in team leadership, system design, testing, and complex engineering problem-solving.

13 · Future Work

Future development could focus on:

• Introducing AI-based predictive models to identify fatigue before it becomes clearly observable

• Adding blood oxygen and skin-temperature monitoring

• Improving ECG quality under movement and vibration

• Miniaturizing the hardware and increasing system integration

• Testing the system under a wider range of driving conditions

• Developing personalized physiological baselines

• Integrating the system with commercial vehicle platforms