Kalman Filter Approach for Mobile Robot Localization Fusing Ultrasonic Sensor and Laser Range Finder Measurements

Abstract

Mobile robot localization has attracted significant attention from the researchers and robotic engineers in the past few decades. To reliably navigate in an environment, a mobile robot must know where it is. Many robotic applications, such as planetary exploration, rescue, and surveillance etc., requires efficient robot localization within unknown environments. This issue involves building a map of the work space and localizing itself with in such map. There are varieties of complex environment where robot need to navigate. One of the issue of localization is to find the location of the robot in partially or completely unknown environment, which makes the problem more difficult to solve. Moreover, physical platforms and sensor suites vary significantly from system to system and complicating the task of sensing and control. In order to increase the accuracy when robot navigates into partially or completely unknown environment, it is necessary to integrate sensing information from distinct sensors for detecting objects and translate the different sensory inputs into reliable robot position estimates with respect to the environment. We propose the approach for mobile robot localization in a partially unknown indoor environment. The propose approach use Unscented Kalman filter, ultrasonic sensor system and laser range finder. We implement UKF localization algorithm fusing the measurements of ultrasonic sensor and laser range finder. This algorithm relies on calculating correspondence of data association for beacons. The data association is computed through maximum likelihood estimator which determines the most likely value of the correspondence variable and the consider it as a true value. While the sensor fusion in this context is for integrating the data from ultrasonic sensor and laser range finder to improve the robot localization.