WalkGuard

Inspiration

WalkGuard was inspired by the need to empower the visually impaired with enhanced mobility, safety, and independence. Seeing the challenges faced with traditional walking sticks, we sought to integrate modern technology—object recognition, obstacle detection, and real-time navigation—into a single, smart device that improves their quality of life.

What it does

WalkGuard is a smart walking stick equipped with:

Ultrasonic Obstacle Detection: Helps detect nearby objects and obstacles, guiding the user around them. YOLO V7 Object Recognition: Recognizes and classifies surrounding objects, providing audio feedback to the user for safer navigation. GPS Navigation: Offers turn-by-turn directions to help users travel independently, even in unfamiliar areas. SOS Button: Sends an emergency alert with real-time location data to predefined contacts in case of distress.

How we built it

We combined hardware and software to create WalkGuard:

Hardware: We used ultrasonic sensors for obstacle detection, a camera for video input, and a GPS module for navigation. A microcontroller unit (MCU) ties everything together. Software: YOLO V7, a deep learning model, powers the object recognition system. We integrated this with custom voice output to guide the user. GPS navigation was handled via an API for real-time tracking, and the SOS function uses communication protocols to send emergency alerts.

Challenges we ran into

Object Recognition Tuning: Ensuring that YOLO V7 accurately recognized objects in varying light and weather conditions was challenging. Battery Optimization: Powering multiple sensors and processing units while maintaining long battery life required careful optimization. Real-time Feedback: Providing instant alerts without delay was a technical hurdle, particularly when integrating GPS and object recognition data streams.

Accomplishments that we're proud of

Successfully integrating multiple technologies into one cohesive system was a major achievement. Achieving high accuracy with obstacle detection and object recognition in real-world scenarios. Developing a reliable SOS feature that works seamlessly in emergency situations.

What we learned

We learned how to integrate machine learning models like YOLO V7 into real-time applications and optimized hardware for low-power operation without compromising performance. Additionally, we gained valuable experience in user-centered design, focusing on ease of use and accessibility.

What's next for WalkGuard

We plan to:

Refine object recognition further for more diverse environments. Improve the user interface, including voice commands and feedback. Explore partnerships to bring WalkGuard to the market and impact as many lives as possible.