Problem :
Blind individuals encounter challenges when participating in swimming activities because they lack a reliable navigation system tailored to their specific needs. Existing solutions fall short in providing real-time guidance, hindering visually impaired swimmers from confidently navigating pools independently. This limitation compromises their safety and restricts their access to the numerous physical and mental health benefits associated with swimming. There is a pressing need for an innovative solution that addresses these challenges, ensuring that blind swimmers can enjoy a more inclusive and empowering aquatic experience.
Solution :
I am introducing my innovative AI-powered swimming navigation system, a game-changer for blind swimmers seeking an inclusive and independent aquatic experience. This groundbreaking solution combines a camera, Raspberry Pi, and three haptic motors to provide real-time guidance in the pool. Users receive haptic feedback for directional corrections and distance notifications, enabling them to swim confidently and safely. This AI-powered machine eliminates barriers to aquatic recreation for visually impaired individuals, promoting water safety and ensuring an enriching and accessible swimming experience for all.
Result :
By the end of this implementation, you'll have a fully functional AI-driven swimming navigation system, providing visually impaired individuals with the ability to swim independently and confidently. This is a great solution because it's affordable, durable, and reliable! One of the best aspects is the "leave and forget" ... it's fully autonomous, so you won't have to stand and configure it for 10 minutes - just turn it on and swim! Dive into a new era of accessible swimming with my AI-driven machine.
Get ready to build your project! Gather all the necessary parts - Raspberry Pi 4, power bank, SD card, 3x haptic motors (3 volts), and a webcam.
If you're using Raspberry Pi for the first time, I recommend watching this fun video on what is a raspberry pi! Now get your Raspberry Pi, install the Raspian OS on the SD card, and boot it! Here's a tutorial on how to do it.
After doing that, go to the terminal and update the software. Use commands (write each line separately) :
sudo apt-get update
sudo apt-get upgrade
sudo restart
Now let's install all the necessary packages needed, to run this project :
pip install teachable-machine
pip install opencv-python
git clone https://github.com/letmaik/pyvirtualcam --recursive
cd pyvirtualcam
pip install .
pip install RPi.GPIO
If, while installing, the Raspberry Pi gives you system packages error, after the pip lines write :
--break-system-packages
Example :
pip install teachable-machine --break-system-packages
Hardware assemblyNow take all your parts and solder them together as shown in the picture!
After hardware assembly, you just have to download the code to your Raspberry Pi! I've already trained 2 AI models, that measure the distance and keep you in a straight line.
All you have to do is download my GitHub repository. Here's how:
git clone https://github.com/CMPL3X/Helping-People-W-Visual-Impairments-Swim.git
Before running the Main.py code, run cameraPortTest.py. It will show you on what port your webcam is working. If the window with your camera is titled "port 0", you don't have to change anything, but if it's something else, go into the Main.py and change the line :
camera = cv2.VideoCapture(0)
to whatever port you have.
Now, all you have to do to run the code is write :
cd Helping-People-W-Visual-Impairments-Swim
python3 Main.py
To automatically start the code, when the Raspberry is turned on, use this tutorial.
Now all you have to do is just make an enclosure, for this project. You can be as creative, as you want. Suggestion - you can also add a wireless charging power bank and you won't have to worry about opening the device to charge it.
My solution - I've added 2 3d models (below) that are already made enclosures. It consists of a top part and a bottom part. You can put everything in, glue them together, and drill a small hole for the wires that you can just glue (for example, with epoxid) together. You can put the camera, Raspberry Pi, and power bank in there. I suggest buying an organic glass plate (they're cheap) and cutting it down to size. Then you can slide it in and guide it. At the bottom, there is a hole for a belt. Then you need to take motor wires and get them to the glasses. The motors have their tape, but I recommend using double-sided tape. You can glue them in the glasses and tie together the wires and that's it!
For this device to work you have to put 2 "X" in each underwater wall at the end, in the middle. The AI will recognize it and correct your swim path, based on it! I suggest screwing this. Or make a locking magnet/suction cup mechanism (If you're not the pool owner).
Comments