SeatSense

Problem

Every university student will have that experience when they enter a busy library, they cannot seem to find a free space, so they are running up and down the entire library just to find an empty seat to work in. You think of checking for an available seat but the university library website only provides you with the information about reservable seats and not FCFS seats. This is a waste of valuable time and energy, especially for university students with tight schedules. So, we came up with the solution - SeatSense - a system that saves valuable time for university students studying in libraries. The web interface also allows students outside the library to know how busy the library is so they can judge whether there is a space for them in the library.

Inspiration

The inspiration came from auto checkouts at local supermarkets like Tesco, or Sainsbury's and also from underground car parks in huge malls. They both have a system where the LED light turns green when the spot is empty and free to use and red when the space is occupied. We took that idea further and applied it to university libraries.

What it does

Ultrasonic sensors and video cameras are used in conjunction to detect user presence in the seat. When the seat is empty the LED turns green and when the seat is occupied the LED turns off so the user is not disturbed. This availability status is updated to the university library website which enables the user to look at the website for available seats without needing to adventure through the entire library to physically search for free seats. The website has seat layouts identical to the actual library for feasible comprehension of the map for the users.

How we built it

We used lovable to develop the API and the website. We used an esp32 along with an ultrasonic sensor and the Arduino editor to create the ultrasonic detector. For computer vision we used Kotlin on Android Studio and the yolo26n model to detect the presence of people.

Challenges we ran into

A few hours were wasted trying to use the Arduino Uno Q because the drivers were not updated and gave us issues interfacing with the device. Also, linking the Arduino circuit to the demo website we built proved a challenge in which we invested a lot of time on.

Accomplishments that we're proud of

We successfully created an interactive demo system with a mixture of hardware and software that accurately detects humans via ultrasonic sensors and image detection and updates the LED light and the website accordingly.

What we learned

We learnt that understanding hardware components and learning the appropriate applications for each components is very crucial. Also, although gen AI is reducing a lot of workload for us, it is vital that the developers are also aware of the full technology so they can respond to any bugs and errors appropriately.

What's next for SeatSense

Improving the YOLO model detection. Make the API secure. Making the method of detection more reliable. Polishing the system so the LED is more visible for people and cameras and sensors are hidden out of sight for aesthetics.