Localization of a mobile robot using ZigBee based optimization techniques

摘要

This paper introduces techniques for designing and implementing remotely controllable outdoor robots with the ability to localize itself while traversing in hazardous environments. The robot platform designed is a semi-autonomous tiller modified as required by the application. Localization based on wireless sensor network which has been a research interest of is addressed here with the use of Zigbee protocol. A Kalman filter was applied for the initial gathering of training data. For the improvement of the accuracy of position estimation, a fuzzy inference system is introduced. An extended Kalman filter is applied at the end in order to fuse the encoder data with the filtered RSS data. Results prove that a metric accuracy varying from 2m to 10m is achievable depending on the location of the robot with respect to the sensor nodes.