The Effect of Number of Landmarks in Memory Management Methods on Simulation of Mobile Robot’s Simultaneous Localization and Map building

In this paper we discuss techniques to add memorial methods to a simulation implementation of robot running Simultaneous Localization and Mapbuilding (SLAM) in navigation and exploration by mobile robot systems. Two main methods are: First method is a solution to the relocation problem usinggeometric constraints, joint compatibility, and random sampling in filter, and the other method is dynamic memory characteristic including both a limited numberof feature storage as well as a feature decay function. These innovations methods and processes, making the robot more robust, capable of handling large featureladenenvironments, as well as the relocation problem. We demonstrate simulationally in Matlab the benefits of these methods with experiments in a simulated environment

کلیدواژه ها:

SLAM ، Localization ، Memory ، mapping ، Simulation