Vision based navigational approach for controlling of mobile robots

Abstract

Navigation plays an essential and important role in robotics. When the mobility is embedded to a robot the navigation is a must. Most of the robots are using close range mono-feel sensors for their navigation purposes but it is not reliable for the most of real world purposes. Computer vision is a major research area of the computer science. Navigation using vision is very accurate, applicable and reliable. It reduces the usage of unnecessary sensors. Active vision and image processing theories are applied to identify lines, curves and objects. Robots can be trained by storing specific objects on the path in an image database and SURF algorithm can identify specific objects in the real world by referring the database. OpenCV will provide a better approach for image processing. Localizing the robot to object is according to the vision map and can be done using PID algorithm and fuzzy logic. Fuzzy logic will be applied for making of an accurate movement. This paper will discuss the active vision object finding using applicable image processing algorithms to control the mobile robots in accurate manner.