The Air hockey project is basically an autonomous air hockey platform involving a serial manipulator that can play air hockey with a human competitor.
The autonomous air hockey platform involves design and fabrication of an air hockey table and a 2 link serial manipulator to control the mallet. The table consists of 2 levels, the lower one to host the blower fans and the upper one acts as the table surface to play the game The location and movement of the puck is perceived using image processing algorithms from the real time video feedback from a camera placed above the table. Based on the obtained values, the precise trajectory of the puck is found. Based on the calculated position at the given time, the motors actuating the links are given predetermined angles and speeds using inverse kinematics algorithms. Thus, the end effector hits the puck with the required force.
The future prospects include using machine learning algorithms to improve tracking of the puck in diverse conditions and to reduce the error in trajectory calculations and inverse kinematics algorithms and faster response times.