How did you feel when you first sat behind the driving wheel? Scared? Excited? All of us on the team felt a similar way: nervous. Nervous that we'll drive too slow and have cars honk at us from behind. Or nervous that we'll crash into something or someone. We felt that this was something that most people encountered, and given the current technology and opportunity, this was the perfect chance to create a solution that can help inexperienced drivers.
What it does
Drovo records average speed and composite jerk (the first derivative of acceleration with respect to time) over the course of a driver's trip. From this data, it determines a driving grade based on the results of a SVM machine learning model.
How I built it
The technology making up Drovo can be summarized in three core components: the Android app, machine learning model, and Ford head unit. Interaction can start from either the Android app or Ford head unit. Once a trip is started, the Android app will compile data from its own accelerometer and multiple features from the Ford head unit which it will feed to a SVM machine learning model. The results of the analysis will be summarized with a single driving letter grade which will be read out to the user, surfaced to the head unit, and shown on the device.
Challenges I ran into
Much of the hackathon was spent learning how to properly integrate our Android app and machine learning model with the Ford head unit via smart device link. This led to multiple challenges along the way such as figuring out how to properly communicate from the main Android activity to the smart device link service and from the service to the head unit via RPC.
Accomplishments that I'm proud of
We are proud that we were able to make a fully connected user experience that enables interaction from multiple user interfaces such as the phone, Ford head unit, or voice.
What I learned
We learned how to work with smart device link, various new Android techniques, and vehicle infotainment systems.
What's next for Drovo
We think that Drovo should be more than just a one time measurement of driving skills. We are thinking of keeping track of your previous trips to see how your driving skills have changed over time. We would also like to return the vehicle data we analyzed to highlight specific periods of bad driving.
Beyond that, we think Drovo could be a great incentive for teenage drivers to be proud of good driving. By implementing a social leaderboard, users can see their friends' driving grades, which will in turn motivate them to increase their own driving skills.