Inspiration
We got inspired by some memories of a flash game we used to play. It was supposed to help us practice simple sums, all while driving a jet ski.
What it does
The project can be divided in two parts: the frontend and the backend. The frontend chooses a random image and asks the player what element is in the image. If they respond correctly, they get a point and advance one position. If they don't, they'll lose a point and recede one position. The backend only needs to be run sporadically, and it generates new levels. It does so by scraping a Google Images triggered by a determined search term, and sends the image to the RestbAI, which detects features and appliances and selects randomly one of the ones with the highest confidence. These are then intertwined with similar answers and prepared to be sent to the user when needed.
How we built it
To build it, we divided the process in 4 parts: designing, building the backend, building the frontend and building the mobile app. We used NodeJS for the backend, ReactJS for the frontend and Cordova for the mobile app.
Challenges we ran into
Apart from stupid bugs we found from time to time (like forgetting we had 50 elements and not 200), we had to find our way around some problems with the REST wrapper we were using. We ended up not using that, opting instead to use Curl.
Accomplishments that we're proud of
Figuring out almost everything by ourselves was very rewarding, as well as finishing the project on time.
What we learned
The two team members that worked on the code learned new skills. Giovanni learned Cordova, and Pol learned how to make and deploy a test app with the HMS toolkit. We also learned that we are pretty resistant against a lack of sleep.
What's next for AI on wheels
We would like to improve the design, and integrate it into a website as a 404 page game. Also, with more time, the database of possible levels can be improved.
Log in or sign up for Devpost to join the conversation.