AlertDrive Pro

Inspiration

The sobering statistics of over 46,000 lives lost annually in car crashes in the United States alone, with a notable contribution from drowsy driving, served as a stark motivator. Research suggests that drowsy driving is responsible for approximately 20% of all car accidents, underscoring the critical need for a solution that can mitigate these preventable incidents. AlertDrive Pro was conceived from the urgent need to address these preventable causes of accidents, aiming to save lives by ensuring drivers remain alert, attentive, and secure through innovative technology. Inspired by the need to enhance road safety, we set out to address the significant contributors to accidents: drowsy driving, hands-off-the-wheel situations, and deceptive seat belt usage.

What it does

AlertDrive Pro introduces an advanced safety mechanism that integrates three critical features: Drowsy Driver Detection: Utilizes facial recognition to monitor the driver’s eyes. If the driver's eyes close, indicating potential drowsiness, the system automatically initiates a wake-up call to the driver's phone. Hand-on-Wheel Detection: Ensures at least one hand is on the steering wheel. If both hands are detected off the wheel, it triggers a phone call alert to remind the driver to maintain control. Seat Belt Detection: Detects whether the seat belt is properly fastened across the driver's chest, addressing the issue of drivers bypassing the seat belt alarm by improperly securing the belt. These calls persist until the driver responds, ensuring their attention is redirected back to the task of driving safely.

How we built it

Over the past 48 hours at HackIllinois, our team embarked on a rigorous development sprint to bring AlertDrive Pro to life. Starting with a brainstorming session to identify the most effective approach, we then divided tasks to parallelize the workload. Our project is built on a foundation of Python, OpenCV, dlib, Twilio, mediapipe, yolov3, etc. Facial landmark detection, eye aspect ratio calculation, and real-time video stream processing form the core of our drowsiness detection system. Hands-on-the-wheel detection is achieved through continuous monitoring of hand positions. Seat belt detection employs innovative methods to differentiate genuine usage from deceptive practices. The integration of Twilio enables real-time phone call alerts for immediate driver response. We developed a prototype that could reliably detect the key indicators of drowsy or inattentive driving. Through iterative testing and refinement, we integrated the system to work seamlessly, ensuring real-time detection and accurate response.

Challenges we ran into

The development of AlertDrive Pro was a journey filled with challenges, including refining object recognition to accurately identify eyes, hands, and seat belts within video frames. Navigating the complexities of object recognition within video frames presented challenges, particularly in identifying crucial facial features like eyes. We faced initial difficulties integrating audio elements for alerts, leading us to adopt phone call alerts for consistent and effective wake-up calls. Additionally, accurately detecting the absence of hands on the steering wheel and distinguishing between temporary adjustments versus actual drowsiness posed significant technical hurdles. Through a combination of algorithmic adjustments and sensor calibration, we managed to overcome these challenges, significantly improving detection accuracy and reliability.

Accomplishments that we're proud of

As a team of second-year students entering our first hackathon, creating AlertDrive Pro is a significant accomplishment. Overcoming challenges in programming, facial recognition, and integrating Twilio services, we successfully developed a functional system that has the potential to save lives. We are particularly proud of our system's ability to function accurately in a wide range of driving conditions, including at night. The development of a non-intrusive, real-time monitoring system that respects the driver's privacy while ensuring their safety is a significant achievement. Moreover, the persistence mechanism of the alert system, ensuring the driver's engagement by requiring them to answer the phone call, represents an innovative approach to addressing drowsy driving.

What we learned

Throughout this project, we've gained invaluable experience in sensor integration, real-time data processing, and integrating telecommunication services. We've gained valuable insights into the complexities of facial and object detection and the critical role of responsiveness in safety technologies. We learned the importance of user-centric design, especially in safety applications. This project also highlighted the significance of team collaboration and effective problem-solving under time constraints.

What's next for AlertDrive Pro

Looking ahead, we plan to further refine Drowsy Driver by incorporating feedback mechanisms for the driver to provide real-time feedback on alerts, enabling machine learning algorithms to adapt and personalize the system's sensitivity and responses. Additionally, we aim to explore partnerships with car manufacturers and ride-sharing companies to integrate this technology directly into vehicles, broadening our impact on road safety. AlertDrive Pro is a versatile platform with continuous opportunities for expansion, and we're excited to bring it closer to real-world implementation.