EVA: The Emergency Vehicle Alert

Website UI

Inspiration

We felt that, because of the diversity in our group, it would make more sense to follow the second track of this project -- focusing on enhancing transportation safety, efficiency, and sustainability. Our group is comprised of two Computer Science majors, two Computer Science and Engineering majors, one Management and Information Systems major, and one Civil, Environmental, and Sustainable Engineering major, so this prompt seemed much more fitting. Once we decided this, we looked to things that affect us in our everyday lives. An idea that arouse was awareness of EMS vehicles on the road. We identified two main concept to address. The efficacy of response times and the safety of the surrounding drivers. These, we felt, we could tackle with AI.

What it does

Our product is a user-friendly website that will use AI to identify if an ambulance is in an image. All the user needs to do is copy the url of an image and paste it on the website. Our AI will then tell the user if an ambulance is present in the image.

How we built it

We brought this idea to life using a combination of React(CSS & JavaScript/HTML) and Python with Flask. First we had two groups of people working. One on the web design and the other on the software design. The backend group was in charge of writing the code, then they used Kaggle to find data sets and train the AI to identify ambulance. The front created the website while also working on the product idea as a whole. Once we had progressed to a certain point, we were able to join up to create the API that connected the backend code to the front end display -- creating a fully functioning design.

Challenges we ran into

We are all first year coders with little knowledge and no AI experience. We struggled to figure out how to use AWS, so we ended up only using Bedrock to create a logo. We also were unsure where to start the whole process, so we scaled back our original plan to ensure we could figure this out, step by step.

Accomplishments that we're proud of

We are proud to be able to present a final idea and code to back it up. At one point, we were unsure if we would have a project to turn in. Now we have created a pretty cool idea plus a fun website to back it up. The site completes the task it is intended to and the python code functions as well.

What we learned

We learned:

Git & Github
How to navigate AWS
Creating a webpage with React
APIs
Collaboration

What's next for Emergency Vehicle Alert

This AI project could become an app that spawns an alert on mobile devices, via notification, when an emergency vehicle is expected to cross paths with a driver. It will give cars advanced notice of EMS vehicles in their vicinity, accelerating emergency response times. Rush hour is hectic, especially in locations with high population density such as the Bay Area. If roads are crowded, emergency response vehicles may have a difficult time reaching emergency situations in a timely manner. Another side of this is the user's experience. We as a society have put each individual human in charge of a metal bullet that can go at speeds up to 100 mph. Especially at these faster speeds, driving can be a stressful environment. When there is a screeching siren and flashing lights coming up behind you, there is added stress. By alerting the drivers beforehand, they are prepared to get out of the way for EMS vehicles. This could be further developed using Yolo as an AI model because of its efficiency when it comes to speed and recognizing large objects. The AI would recognize the ambulances or fire trucks in INRIX traffic camera footage. It would then send a notification through a partner app to alert drivers on the road that there is an ambulance in the vicinity.