A Relative Localization Method Based on Encoders and Micro-inertial System

Abstract

Odometer is used to measure the distance that the wheels rolled when the robot is moving and can be replaced by the motor encoders. But when the wheel comes to idling or skid, there will be errors. Idling or skid generally occurs in the turning process where the information of inertia sensors is required to correct the localization. In this paper, we discuss a relative localization method to keep small error during a relatively long time without using external sensors. The initial alignment is used to reduce the error caused by inclination of the ground. The Kalman filter algorithm which is used to fuse the odometer and micro-inertial system information are proposed to reduce the relative positioning error.