Detection, Recognition, and Grasping of Objects through Artificial Intelligence Using a Robotic Hand

Abstract

In this paper, a robotic hand can perceive surrounding objects and grasp the object of interest. This is achieved by the implementation of the YOLO algorithm for the detection of everyday objects, filtering information on the presence of man-made objects, and sending that real-time information by serial communication on the robotic hand microcontroller. Upon receiving this information, commands are sent to the robotic hand motors and the latter is compressed and takes the form of the object detected by the algorithm. For objects such as bottles and cups, the adequate shape is considered the tightening of 4 fingers except for the thumb, while for thin objects such as forks, spoons, etc. the thumb and index finger are activated.