Object Recognizing Robot Application with Deep Learning

Abstract

Today, technological devices and robots are used to benefit in many different areas. In environments where people of other military field may be at risk, vital risks are desired to be minimized for robots. It is very risky for a person to enter a building for reconnaissance purposes during military operations. It is the place where this kind of risky person learns a robot that can be remotely controlled, recognize the text he sees, and display the text control text he knows, instead of making exploration. In this structure, the robot uses the TensorFlow deep learning library offered by Google to recognize objects. It was run on the Raspberry Pi3/B minicomputer on the software with the language of Python. DC motors are used for robot movement. In Raspberry Pi3/B minicomputer, the robot's movements can be controlled by sending a signal to the motor driver circuit with GPIO pins. In the tests of the prototype, it has been observed that the guarantee of success in the distribution area has increased.