GymQueue

Inspiration

No matter what time of day it is, the Rutgers gyms will always be busy. Enough machines are not there compared to the number of students who want to work out and stay healthy. The biggest issues in the gym arise when there are long wait times for machines. Sometimes, you try to wait for a machine and while you're talking to your friend, that machine just gets taken within a blink of an eye.

What it does

Therefore, we developed a smart vision queuing system for the Rutgers gyms. Essentially, it forces Rutgers students to sign up in queues for a machine. All gyms and their machines are listed in the webpage and you have the opportunity to select.

The intriguing part is the low-hassle setup of this process. You first create an account with your NetID and scan your face. The database stores face embeddings rather than your actual face. Once you have logged in, you join a queue, and when it's your turn at the machine all you have to do is sit down and the camera will recognize that you're checked into the machine. If you don't come on time the next person will be queued.

We have considered all factors of the gym such as people wanting to work in with you. Therefore, you can add people to your current rank in the queue and work in with them with the facial recognition allowing both of you to work out on that machine. You have rest periods where you can step off camera to take a break, get water, or talk to your friends, but once that 5 minute timer ends you need to be returned to the camera, otherwise the next person will be permitted. This all is to avoid the issue of random people joining queues because college kids like to play pranks and join from possibly their dorms.

Rep counter

On top of the queuing system, we integrated a rep counter for exercises. Since we only have one camera currently, the rep counter is only used for Shoulder Press but this is something we can expand upon in the future. The rep counter adds to a dashboard on our webpage where you can also edit the weight, number of sets, and reps. It also properly accounts for people working in with each other.

What we learned

We learned about computer vision, pre-trained facial recognition models, Eleven Labs, and integrating hardware with a web server.

Challenges we ran into

The main challenges we faced was that this was a big project and there were many parts that we needed to coordinate with each other to ensure proper functionality. We overcame these problems through effective communication and making sure our features don't overlap. We also went through many stages of testing to ensure there were no issues.