The problem Potometer solves

If you could root down vehicular accidents, traffic congestion, and loss of life on roads, you would almost always end up blaming potholes on the road. Statistically, India witnessed 747,361 fatal deaths due to poor road conditions. The intensity of threats posed by potholes is simply disastrous. Metropolitan cities, including Bengaluru, have showcased a dismal performance in tackling problems related to potholes. There’s no existing technology that determines the quality of the road based on the pothole density.
Our problem statement is to develop a Machine Learning Algorithm to determine road quality index using a digital motion processor and an integrated GPS unit. This information can be used to suggest appropriate paths to drivers for a better driving experience and lesser congestion.
Further, the project has been taken ahead by incorporating Computer Vision for detecting potholes from images. This increases the accuracy of detecting potholes by using numerous filters and a convolutional neural network.

Challenges we ran into

One of the challenges we faced was that the GPS unit would not work indoors. We had to sit near the window to get satellites to connect to so that we can test our model.
We had issues passing props to child components in React since we were using function components. We resolved it by changing them into classes and using state.
The quality of documentation of react google maps api was not particularly legible. We had to refer multiple sources to solve the issues we faced.
The motion sensor doesn’t distinguish between potholes and speed bumps. This was resolved by implementing CV along with the hardware.