Our team decided to dedicate our project to solve a problem for people with disabilities. The inspiration came after realizing how often certain disabilities can be overlooked. Our motivation for this project was enhanced further after meeting a blind student named Justin Heard on Friday afternoon, only hours before the hackathon began. After he explained a blind persons toughest challenge is to detect overhead objects, we knew exactly what we needed to do to solve that problem.
What it does
WAND is a device that is able to detect and identify objects that may pose a risk to a blind person walking. The WAND has a camera for imaging input. The images are then run through OpenCV for object detection which returns a word representation of the objects in the image. The word of the center most object is then put through a text-to-speech conversion which outputs audio to the blind person using it. The ultrasonic sensor is able to tell the distance of the object, so the combined output to the user would be something like, "Chair 5ft". WAND is also an acronym that stands for Walking Aid Notification Device.
How we built it
We started by hooking up a button, camera, and ultrasonic sensor to a raspberry pi. The camera gets activated by a button, with an ultrasonic sensor to figure out how far that object is. We use object detection and text-to-speech to have the object and distance output with audio.
Challenges we ran into
A big challenge our team faced was getting our hardware issues diagnosed. There were moments where we thought that we may need to find a different camera and/or sensors because they were not working as expected. We were tempted to go out of the way to go buy a Raspberry Pi camera, but with more in-depth research, we figured out how to get the logitech camera functioning properly. Importing libraries to the raspberry pi was difficult because we were quickly trying to test various methods of object detection. We were constantly importing new libraries for every method that we were testing and eventually we hit the max memory capacity, so we had to quickly figure out how to reallocate the partitioned memory and ensure that the device wouldn't run out of memory during demonstration. Connecting pins on the circuit had to be double and triple checked over and over. After looking at a circuit for too long it is common that a person may not be able to identify mistakes that they have made, so we made sure to rotate fresh eyes to diagnose circuit connections quickly. The first ultra sonic sensor we used stopped working suddenly, and initially we thought this was a software problem. We spent many hours of our last day unit testing and rewriting all of our software over again but then we ultimately found out that our ultrasonic sensor burned out the day before, so the problem was solved by simply replacing it with a new one.
Accomplishments that we're proud of
Having a member of the blind community tell us that our concept is "unique" in comparison to other devices that have been tried over the past decade was very rewarding. We were told that the toughest challenge for a blind person is the ability to detect over head objects that are missed by a standard walking stick. Overhead objects can cause serious concussions and damage to a blind person's head walking at normal speed. So the purpose of our project was to create something that would give a blind person the ability to detect and identify objects in any direction even if they are out of reach. We are proud to be able to address one of the toughest challenges facing the blind community today, and being able to get input from a member of the blind community helped us understand and solve a unique problem in an intuitive way.
What we learned
We learned a lot about the tough challenges facing the blind community and gained the perspective needed to address these issues correctly using our specific expertise. We learned to focus on our individual strengths in order to effectively accomplish many tasks in a timely manner. We also learned a lot more about the skills we have been developing in our course work over the last few years here at UGA.
What's next for WAND
Include vibration haptics to communicate environment surroundings to a blind person. Be able to collect more data to analyze and the surrounding environment in higher detail to improve navigation. Being able to communicate a more detailed environment to a blind person is crucial but challenging.