Humanlike Robotic Vision System Using Artificial Neural Network

Robotic manipulator, which is a kind of robot that interacts with its environment in order to manipulate objects, is the most extensively-employed kind of robot in practical field. Most of the robotic manipulators functioning in the industries and other fields are programmed to manipulate objects from fixed predefined points in space. However, this might create problems in real world situation where the location of the object is arbitrary and cannot be predefined. This problem can be solved if the robot is able to recognize the object by learning which one is the specific object and which one is not. The robot can then find the position where the object is located in order to drive the manipulator tool point towards that position. In this paper, an approach has been taken to design and analyze performance of a humanlike robotic vision system for robotic manipulator using artificial neural network along with image processing. Here, the robot captures image of the environment, analyzes it and determines the points in space where the object to be manipulated is locate.