A framework for real time indoor robot navigation using Monte Carlo Localization and ORB feature detection

This paper has introduced a framework for indoor navigation implemented by using a computer, Android device and Lego Mindstorms NXT robot. The Lego Mindstorms NXT robot explores and navigates autonomously through a known environment, making its own decisions. An Android device is used for object recognition. The robot is able to localize itself based on the landmark observed using ORB (oriented fast rotated brief) feature detection and the sensory data from ultrasonic sensor using Monte Carlo Localization. The robot is able to plan its own path towards the goal using the A* shortest path. The navigation system is able to identify and recognize the landmarks and environment; and reacts accordingly to achieve the goal. Experimental results show that the robot navigation system is successfully designed and implemented with an accuracy of ±38 cm root mean squared error. © 2014 IEEE.