Research On Kalman Filter Based Localization Of Mobile Robot

Robot Localization is a key part of robot navigation,and is the first step to complete mobile task.The main point of this thesis is to investgate the autonomous localization technology in known or unknown environment,The physical experiment platform is TurtleBot controlling by Matlab or ROS.The aim of this thesis is to build a reliable and effective algorithem to solve localization,which is suitable to TurtleBot The parameters using in this algorithem are obtained from experiments.Doing literature review is the first step.Based on the investigation into the literature,the research plan is made.Making experiments with physical robot is to determine some important parameters.And with the theroy of extended Kalman filter,the simulation of dead-reckoning using a single odometry is achieved.With multi-sensors,we focus on the map building and the map-based localization.Whar's more,we have an insight into the principel of Monte-Carlo localization,and carry out experiments on the simulation robot and physical one.The main content of this thesis is:With UMBmark tests,we can find out important parameters to offset the systematic errors caused by the physical system.It can imporve the accuracy of the localization results.Extended Kalman filter was comprensively investigated.According to this method,the robot is able to achieve localization using information from a single sensor or multi-sensors.With the help of map,localization can be more effective and reliable.The artificial features in the environment can be used to build map based on the extended Kalman filter.Meanwhile,the map provides the localization of these features,which is useful for robot to locate its relative position.The principle of Monte-Carlo localization is investigated.The simulation and experiments of this method are achieved on Matlab and ROS platform perspectively.