Vision-Based Mobile Robot Self-localization and Mapping System for Indoor Environment

Abstract

Localizing accurately and building map of an environment concurrently is a key factor of a mobile robot system. In this system, the robot makes localization and mapping with artificial landmarks and map-based system. It is a process by which a mobile robot can build a map of an environment while continuously determining the location of the robot within the map. The system estimates the robot position in indoor environments using sensors; a camera, three ultrasonic sensors and encoders. The main contribution of this paper is to reduce computational time and improve mapping with map-based system. The self-localization of mobile robot in an indoor environment is advanced through the construction of map based on sensors and recognition of artificial landmarks. Vision based localization system can benefit from using with ultrasonic sensors. From captured images, the system makes landmark detection by using Canny edge detection and Chain-code Approximation algorithms to represent the contour of landmarks by using edge points. The Kalman filter is aimed to accurately estimate position and orientation of the robot using relative distances to walls or artificial landmarks in environments. A robotic system is capable of mapping in an indoor environment and localizing with respect to the map, in real time, using artificial landmarks and sensors.