Automatic Detection of Driving Distractions

Introduction

This project, developed for the "Embedded System Modelling" course, implements an embedded system for the Automatic Detection of Driving Distractions. The primary goal is to enhance road safety by identifying signs of driver inattention through various sensors and responding with appropriate alerts or actions to re-engage the driver and mitigate potential risks.

System Architecture

The system is comprised of sensors for data acquisition, a central processing unit for analysis, and actuators for driver feedback and intervention.

Sensors

• Gyroscope: Detects head inclination on two axes to monitor for signs of drowsiness or distraction.
• Steering Wheel Rotation Sensor: Measures steering wheel angle to detect erratic driving or sudden corrections.
• Speedometer: Provides current vehicle speed for context in other detections (e.g., safety distance).
• Distance Sensor: Monitors safe following distance to the preceding vehicle.

Actuators

• Warning Lights: Two LEDs (yellow and red) to indicate varying levels of risk.
• Display: A screen to present data and alerts.
• Automatic Brake Activation: System capable of applying brakes at different intensity levels.

Processing Unit

A Raspberry Pi serves as the central processing unit, running the C code responsible for reading sensor data, implementing detection algorithms, and triggering actuators.

Symptom Detection Logic

The system analyzes sensor data to identify distraction symptoms based on specific thresholds and polling intervals.

Symptom	Sensor(s) Used	Detection Criteria
Head Tilt	Gyroscope	X-axis or Y-axis tilt > ±30° for at least two consecutive readings.
Safety Distance	Distance Sensor, Speedometer	The actual distance is compared to a recommended safety distance calculated from speed. Three risk levels are defined: INSECURE, IMPRUDENT, and DANGER COLLISION.
Sharp Turn	Steering Wheel Sensor, Speedometer	Difference > 20° between two consecutive steering wheel readings while speed is > 40 km/h.

System Actions Logic

Every 200ms, the detected symptoms are analyzed to determine the necessary actions, creating a graduated alert system.

Condition Triggering Action	Action(s) Undertaken
(S1 or S2) and D0	Turn on Yellow Light
(S1 and S2) and D0	Turn on Red Light
(S1 or S2) and D1	Turn on Yellow & Red Lights, Activate Brake Level 1
(S1 or S2) and D2	Turn on Yellow & Red Lights, Activate Brake Level 2
D3	Activate Brake Level 3

Conclusion

This project is a practical application of embedded systems to solve a real-world problem. By combining multiple data sources with reactive logic, it endeavors to make our roads safer by proactively detecting and mitigating risks associated with driver distraction.The project's source code is available here: GitHub - Automatic Detection of Driving Distractions.