Tracking Multiple (Three) Pieces

Finally got the image processing algorithm up-to-snuff so it can detect the chess pieces accurately!

In this demo we’re placing 3 pawns from each side and seeing the software react to moves made.

—-
MG

Robot Chess Competition

Chiara robot

Robot Chess Competition

The Chiara robot has the same gripper design as ours.

I guess great minds think alike… 😛

—-
MG

Fall Term Demo

Integration of Chess AI and chess single piece tracker. The graphical representation of the board shows the current state of the chess board after each move. Also, displays the move made by the human and the AI in the cell name format.

Peaceeeee outttt! 🙂

—-
DK

Tracking the Queen and Displaying move

Starting from a blank chessboard, initializing the Chess grid with 0’s (empty squares). After adding Queen, it is recognized and the square value is changed to 1.
The square value changes when the Queen is moved to a new square. The new occupied square now has 1 as the value, implying the square is filled. Also, in the console window, the move made by piece is shown as b2 to b4. For the second test, the queen is moved again to a new position and now the console displays the move b4 to d2.

NOTE:
1) Image processing is very sensitive to light and brightness. Having a vinyl chessboard (highly reflective) makes it more tough to see pieces on the board and get an accurate read.
2) Reading a black piece on a black square and white piece on a white square is dependent on the lighting. (threshold changes due to lighting and is still being worked on)

——
DK

Initial Template Mapping – White Square

Test Successful under dim lighting

—-
DK

Initial Template Mapping with Black square

Black on black hard to read.

—-
DK

Building a Chess Playing Robot Arm

This is a blog created to chronicle the development of the “Human-Robot Chess” project. This is a Senior Design research project being worked on by 4 engineers at Drexel University and their adviser:

• Daksh Kapoor (Computer Engineer)
• Mayank Gureja (Computer Engineer)
• Evan Dissanayake (Electrical/Computer Engineer)
• Anson Jacob (Electrical/Computer Engineer)
• Pramod Abichandani (Adviser)

The purpose of this project is to build a robot arm that can autonomously play chess against a human opponent. This requires the construction of a mechanical arm, an image processing software and a chess algorithm. The project will use a Microsoft Kinect Sensor as the visual input. The sensor will be used to identify the current state of a chess board, pass this information to an AI chess program, and command a robotic-arm to move the chess pieces.

Our final objectives are:

1. Create completely autonomous chess playing robot arm
2. Use Microsoft Kinect sensor to map the chess board
3. Create a robotic arm
4. Integrate chess algorithm, Kinect sensor and the robot arm to play a complete game of chess with a human opponent

In the following months, we will regularly update this blog with our progress.

————————
MG, DK, AJ & ED