Abstract
The usual method for position estimation of a mobile robot is odometry. It has the problem that the estimations errors are accumulated as robot moving and the accuracy of the estimation decreases. To deal with this problem, a positioning technique based on external references is needed. Two approaches are used: the matching between the sensorial information and a map of the environment, or the detection of natural or artificial landmarks. This paper presents a non-linear filter based on a genetic algorithm as an emerging optimization algorithm for mobile robot localization. The sensors used are a camera with a motorized zoom on a pan & tilt platform and a peripheral ring of 24 ultrasonic sensors. An Extended Kalman Filter is used to correct the position and orientation ofthe vehicle. An efficient genetic algorithm in proposed to search for optimal positions. The resulting self-localization module has been integrated successfully in a more complicated navigation system.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
