BlindHelper

A device that gives blind people information about their surroundings to help them with their daily activities.

Comment

Inspiration

I wanted to make something that blind people can wear so that they can carry out their day to day activities with ease. I thought that making a helmet that blind people can wear, in addition to their pre-existing device, would help them navigate their environment easier.

What it does

The helmet has ultrasonic sensors to rapidly detect obstacles in the person's way. If the helmet detects anything less than 50cm away from the person, an alarm will sound, and the LEDs on the helmet will change to red. The LEDs change colors to alert people (if the obstacle detected was a person). The second part of the project was the phone. The helmet also had a spot to place a smartphone in it. This way the smartphone can use its camera to do OCR (Optical Character Recognition) and also analyze the image using image recognization.

How I built it

I first built the helmet using sunglasses and cardboard. Then I glued two ultrasonic sensors to each lens. On the top, I added the LED strip, ESP32, and breadboard. Finally, I added the speaker above the ear. I then programmed the ESP32 to read data from the ultrasonic sensors, process it to determine the states for the LEDs, and made it connect to my computer over WiFi. One of my teammates used the OCR library in android to do character recognization, and my other teammate used YOLO neural network to handle image processing.

Challenges I ran into

When we started making this, I initially used the ESP8266 with the Arduino Uno. This approach didn't work, and for a long time, I spent debugging the serial connection between the two. After hours of debugging, I still couldn't pinpoint the problem, and I decided to use the ESP32 board with integrated WiFi.

Accomplishments that I'm proud of

We were able to get the helmet and OCR working properly to showcase a prototype.

What I learned

I learned more about the Arduino, and how to write custom PWM signals.

How it fared at a HackPHS

Blind Helper did really well at HackPHS. It won first overall, the best use of Google cloud, and best IoT by Ark.io. We are really proud of what we made in 24-hours.

Built With

Share this project:

Updates