Inspiration
Studies estimate between 6,400 and 6,726 fatalities occur annually due to drowsy driving. Over 100,000 to 109,000 people are injured in drowsy driving accidents yearly. A lot of these could have been avoided if there was someone to look after the driver, a co-driver. So, we built something that never gets tired, never looks away, and is always listening.
What it does
It acts as a co-driver to keep checks on the driver, and monitors vehicle speed, velocity, steering movement, and dozens of other data points in real time. AI uses this data to check-in, flag risks, and keeping the driver sharp throughout the trip. Co Driver talks to you naturally using human-quality voice synthesis. You can chat with it to stay awake on long drives or ask it to search for music completely hands-free. Using facial and emotion analysis via the phone's front camera, it detects if you are getting drowsy, distracted, or stressed, intervening with conversation or alerts before an accident happens. In the event of a sudden impact, the hardware sensors trigger a crash detection protocol. Co Driver initiates a 5-second countdown and will automatically simulate a 911 call if the driver is incapacitated.
How we built it
We chose breadboard as our base component. On it, we used ESP32-S3 as the transmitter. It takes the data measured by the sensors APU 6050, INAP 441, and Arducam HM01B0. This data is then integrated with AI in real-time.
Challenges we ran into
We are all Computer Science majors in my team. We don't have a prior experience of working with Arduino or any hardware. We attended the IEEE workshop to learn about it. Used Gemini to learn about it's working. We learned everything from scratch to know how to integrate the hardware with AI. It was very challenging, but very interesting at the same time.
Accomplishments that we're proud of
Learning how to work with hardware is an accomplishment that we're very proud of. We are very proud of working through even though it all seemed very unfamiliar to all of us.
What we learned
First, we learned how hardware works, and then we learned how to integrate AI with it. It was a very great learning journey.
What's next for Co-Driver
We are looking forward to saving a lot of lives and making a positive impact all around the world.
Built With
- amazon-web-services
- antigravity
- arduino
- claude
- esp32-s3
- firebase
- gemini
- github
- google-maps
- inmp441
- mpu6050
- openai
- pip
- python
- pytorch
- react
- sci-kit
- tenserflow
Log in or sign up for Devpost to join the conversation.