Inspiration
Ever felt overloaded with data while driving a car? Struggled with understanding the system? Did not understand what the weird red and yellow flashing symbol on your dashboard means? Why is there a bagel on my dashboard while I'm driving? Did I just leave my car door open overnight? We built DriveWise AI to answer these questions and many more.
What it does
It takes in data from the FordConnect API about the vehicle it is programmed to. A companion flutter app is provided for easy user interface. Natural Language Processing the main idea behind the user interaction keeping in mind that main target user is the person driving the vehicle and they might not be able to access their phone and fumble through the app interface. Also, it is much safer if the person driving the vehicle does not shift their focus from the vehicle and the road to their phones.
It will also keep track of the user's driving habits and perform driver profiling to provide highly customised suggestions to improve the overall driving experience. For example, if the driver usually goes for a long drive over the weekends, the program will suggest the driver to charge the vehicle before the weekend. Once, the driver reaches their home, it will remind them to put the vehicle to charge and also inform them if they left the door unlocked or if the door is not closed properly.
It also has predictive maintenance function which reminds the user to take the vehicle for maintenance or filter change or whatever the need be.
How we built it
The project was built in three layers. The topmost layer is the mobile app that is built using Flutter. The innermost layer is built on FireBase and it takes data by means of FordConnect API. The data is extracted from the database by means of Natural Language Processing using DialogFlow offered by Google Cloud. DialogFlow was chosen as it provides a quick and easy connectivity between the firebase backend and DialogFlow.
Challenges we ran into
Working with the FordConnect app was a challenge due to our limited experience with the APIs. Next challenge was collecting the data and pushing it to firebase.
Accomplishments that we're proud of
We might not have been able to build what we aspired to but we were able to work with DialogFlow and was able to develop an almost working agent.
What we learned
We learnt the working of APIs, especially the FordConnect API and the companion sim. Also about DialogFlow and Firebase. The major learning takeaway is time management.
What's next for DriveWise AI
Finishing the POC and fine tuning the AI to be able to tackle many more use cases and increase its scope of usage. Integrating it with an active working vehicle to test the solution in a real-world setting and enhance its working.
Log in or sign up for Devpost to join the conversation.