EcoDrive

Inspiration

Our team has a shared passion for automobiles, and when we saw Toyota’s challenge we believed we had an innovative solution to solve their problem.

What it does

We designed a device that can be plugged into a car's OBD port to track live data from various sensors on the car. This can give you real-time tracking of fuel economy as well. With this device, there are ports to also add external sensors, such as an outside temperature sensor and incline sensors. This data is stored in an Excel format on an SD card, from which Toyota employees can input the data into our website, EcoDrive. EcoDrive graphs the relationship between measurements made by our device to fuel economy, which can then be directly compared in our website to the the city fuel economy, highway fuel economy, and combined fuel economy of the specific car model from 2021-25. This allows employees to gain a clear understanding of when their car is performing below average, and they can research solutions accordingly. This could be especially important when looking at location as they can develop solutions to niche problems found in certain areas only. EcoDrive also allows employees to view fuel economy ratings of Toyota vehicles from 2021-2025 and even allows a compare function to look at two cars side by side. This can help with predicting future trends.

How we built it

Team RegistrationOur hardware device was coded using an Arduino Uno and a temperature sensor. We also designed a PCB, and the device’s case was created in SolidWorks and includes an SD card slot along with I/O ports. We then developed software that can decode the raw data from the SD card. To parse the Toyota fuel data efficiently, we filtered the data into various fields including model year, manufacturer, transmission, cylinders, and fuel efficiency on different roads. This data was stored in a database using MongoDB for scalability and fast retrieval. For our front end, we used next.js to represent our data using graphs. We also implemented dropdown features for users to easily filter and analyze data. We utilized a REST API to facilitate seamless communication between our MongoDB database and our front end. The database handles specific queries based on the parameters inputted by users, and it is sent to the front end in JSON format for visualization.

Challenges we ran into

In our original plan, we wanted to train our own LLM to act as a chatbot for employees when they have questions about the data. Unfortunately, we were unable to train it ourselves. Instead, we used SambaNova’s API to create our chatbot. Our team also had some difficulties with time management for the first half of our project, as we had some trouble thinking of a meaningful solution. After this delay, we set ourselves hard and soft deadlines, leading to a huge increase in production and lots of progress.

Accomplishments that we're proud of

We’re proud to have created a product that we think will be genuinely helpful for Toyota’s employees. We worked hard to create a functioning full-stack project that even utilizes hardware, and we believe our work has the potential to be a very useful tool for data analysis.

What we learned

We learned how to use new technologies. For all of us, we each used something that we were not as familiar with. For three of us, it was the first time we had ever used hardware. For the electrical engineer, it was the first time he used TypeScript. We learned how to collaborate with new people, as we made our team on the day of.

What's next for EcoDrive

While we were able to implement temperature sensing through our hardware, in the future EcoDrive could be used at a national level, where there would be location based testing, and we would create a graph of the US with fuel economy data all across. This would make it easy for Toyota to see how certain cars perform at certain locations, and use that to make more efficient decisions when designing cars that are more specific in certain locations. Currently, we only analyze the data of Toyota’s cars and compare between its own models, but we want to implement a feature to compare against industry competitors and even have a forecasting tool to see how much fuel costs would be in 5, 10, and 15 years. This could help determine which car has more value at similar price points, and could help Toyota sell vehicles over their competitors if they use that data.