AR-Project-IoT

An Internet-of-things game that models the Star Wars Universe, by prototyping invisible light sabers and visors. Topics include: Localization, Speech Recognition and Motion Classification.

This project was created under the direction of Professor Gregory Pottie and TA John Wu for the design capstone class ECE 180D at UCLA over 6 months and 2 quarters.

Presented in: Annual Research Review | UCLA ECE Department | Spring 2018

Understanding Game Play

1. Two Player Game (D. Vader (Red) and Luke (?) (Blue))
2. Each Player equipped with
   1. 1 IMU (Edison) with custom-designed sabre hilts (C)
   2. 1 VISOR (Edison) with WebCam AND LED Strip attached and configured (Python)
3. Quick Actions performed by either player (asynchronously) involving Stabbing, Blocking and other more complex motions
4. AIM: Must learn the ways of a Jedi and must defeat your opponent in duel. To avoid losing, preserve your health... beat your opponent over 5 rounds of increasing difficulty

How The Game Works

1. Each player exerts an action with the sabre hilt (IMU) and the Edison embedded within it, relays the classified action to server (Yoda?).
2. VISOR determines if the opponent is within HIT-RANGE for that given hilt action (observe the glowing LEDs on the hat).
3. In addition, MIC on VISOR also catches the voice commands intended to increase health points of self during a level. They take practice to perform, and have a defaulted minimum time between two calls.
4. Server receives these signals and has designed game states to determine the health and level (Status) of either player at any point during game play. States traverse 5 levels to find winner. Note: there are sound effects in our final video demonstration.
5. We use WiFi to communicate with Server and our 4 clients (TCP-IP) (C).

See the full game in action here: https://youtu.be/8k3PImhCUhA

The Training Phase

1. To train the duellists, the feedback-based training module is built separately on UNITY (acting as a separate server)
2. The UNITY server classifies mis-actions to train the players to play the game within the scope of the rules.

See the training in action here: https://youtu.be/8w2HDfO5Ouc

Details about specifics in design can be understood in the comments in the code (or you could post an issue). (The contributions in this projects are a reflection of efforts by Donna Branchevsky, Aidan Wilson, Haoran Ma and Pavan Holur, and we have used other open source modules unless otherwise cited)