Investigation on optimized relative localization of a mobile robot using regression analysis

Abstract

The localization of a differential drive wheel robot is a basic control issue in versatile mechanical autonomy. Odometry is a natural technique for deciding the relative localization of mobile robots. It includes the identification of an arrangement of kinematic parameters that allow remaking relative localization of mobile robots and introduction beginning from the encoder wheel data evaluation. Therefore an effective path following technique is desired to develop an autonomous mobile robot with less odometry error. This paper describes a method to identify the optimal value of wheel velocities for relative localization of differential drive wheel robot on a circular and straight path. The proposed technique describes the dependency of odometry error on wheel velocities. Experiments have been performed by employing a two wheel differential drive mobile robot. Linear regression analysis is done to find the relationship between wheel velocity and the odometry of the mobile robot. Confirmation test is also done with the implementation of ANOVA techniques, by using statistical tools available in scientific analyses software MINITAB. A simulation software V-Rep 3.2.1 was used for the validation of the proposed method.