Carrom-Playing-Bot

Abstract

• Carrom is a cool game to play with hand,but on other hand designing a machine to play it is pretty challenging. We have made a bot that can put the coins in the holes by first detecting them through image processing and then deciding which coin to be put in hole i.e most easy to put (by a sophisticated algorithm).
• After the decision is made the mechanical model controlled through aurdino is used to flick the striker.This mechaical model was on a lead screw arrangement for lateral movement along egde of board and the flicker was attached to a stepper motor to shoot the striker.

Mechanical Aspect

• The whole being heavy due to the weight of metal parts, rests on wooden blocks on the frame of the Carrom Board.
• Lnear motion is provided by the lead screw mechanism connected to DC Motor and Encoder.
• The weight of the system is supported by the Linear Rail Rod with Linear Bearing.
• There is two type of rotatory motions, one parallel to board for the angle of hitting and second perpendicular to the board for the purpose of hitting the strikerboth provided by the stepper motors.

Cost Structure

Materials	Cost(INR)
Lead Screw with nut	700
Arduino UNO	500
2 Motor Driver	500
coupling material	250
carrom board	360
Total	2310

Limitations

• If we define 4 regions on the board formed by intersection of two lines joining opposite pockets. If the coin is present in one of the 4 regions farthest from striking position ,straight head on shot will fail.
• Uniform power for all coordinates can lead to unstable movement of coin.
• Rebound shots not taken.
• The radius of the coin is not considered so if a coin is considerably near to the path of the shot it may even obstruct the shot and the result will not be achievable.

Future Improvement

• The current Bot takes shots so as to have head on collision between the striker and coin. If it takes into account the radii of coins and striker it can manage to play oblique shots.
• A further optimization of the current bot is to take rebound shots effictively on the carrom board.
• We can futher include the relation between length and power of shots.
• There is a good possibility of introduction of supervised Machine learning. So the bot improves itself with time.

Team Members

• Akash Deep
• Ayush
• Deovil Vimal Dubey
• Inzamam
• Keshav Dixit
• Ravi Chandra Burra

Mentors

• Yawan
• Dhruv