Supervisor
KP Chan
kpchan@cs.hku.hk
Our Team
William WANG
wyh5625@hku.hk
Project Introduction
AI has become more and more smart in playing chess games. It has even defeated top human players. But it is a mere computer program that can not accomplish physical tasks in the real world. Have you ever thought about playing chess with a AI player in physical world? Great, the idea is about to come true in this project. This project aims to make a robotic arm for it which can move pieces on chessboard with command receiving from AI.
Progress
Milesone: Deliverable 1
- Detailed project plan
- Project web page
- Build robotic arm
- Make a simplifid version of chess app
Milesone: Deliverable 2
- Develop app that can control robotic arm for playing chess
- Apply phone camera so that it can recognize position of chessboard
Milesone: Deliverable 3
- Apply DeepChess model for training
- Import trained AI model onto app
- Prepare presentation and exhibition
Methodology
Working Logic
Chess game is a turn-based game where two players take turns when playing. Therefore, analysing one turn of AI player is sufficient to obtain the overall functions of system.
Robotic Arm Structure
In the first phase, a very smooth and flexible robotic arm will be designed which is controlled by a controller. The preview of completed robotic arm will be like the picture one the left.
DeepChess Model Extract features from position.
Extract features vectors from position, and compare with that of other position
Open CV
Scans chessboard via phone camera, and present the position in the way that computer can understand.
Result
Coming soon...