Olympiknights!

Gameplay
Title Screen

Inspiration

Originally inspired by Just Dance, we saw potential in using motion capture to address accessibility and retention issues seen in modern health apps. We wanted to create something that was engaging alone and with friends.

What it does

Tracking the movement of up to 2 players, Olympiknight's core gameplay loop is a lane based tower defense where different troops are placed depending on different poses and plyometric exercises. This encourages healthy competition between 2 players (which can be done online), or a solo exercise "mission" where quick decision making is required alongside ones workout to win.

How we built it

The game itself runs in Godot's native game engine. Python deals with the heuristic based data from the webcame to determine what moves are being performed by both users. We split work evenly between all the members, pairing up our senior members with beginners to allow learning opportunities for all members of the team.

Challenges we ran into

A primary challenge we ran into were hallucinations in our motion capture technology. Due to the structure of exercises, many specific exercise methods will appear as another to our motion capture screen. We solved this problem by learning to purge unnecessary information nodes, as well as setting a "base" state for a player standing, which would allow us to catch cheaters as well as false positives.

Accomplishments that we're proud of

We are extremely proud of our motion capture system that tracks different body movements, where we spent the majority of our time. Being flexible to changes in colour, light, and speed is something that was a massive undertaking in Python, and we are happy it came to fruition.

What we learned

This was the first hackathon for 2 of our members, and both of us learned a lot about Godot game engine– creating backgrounds and platforms using asset libraries and collisions. Python skills were also improved, learning about how one can send packets of information between the different systems of Godot & Python. Our other members have been to Deltahacks twice before, and learned about using heuristics to track motion capture data, as well as implementation cases of UDP.

What's next for Olympiknights

Further implementation for a variety of exercises, focused on specific workout goals, alongside a story campaign that will keep users more likely finish their plans. These would be focused on a large variety of body types and workout durations as well. Another accessibility change is an ability to use AI modelling to accurately predict the position of ones body outside of the camera frame, allowing for use in smaller settings.