Publication

Abstract : Indoor localization is a vital part of autonomous robots. Obtaining accurate indoor localization is difficult inchallenging indoor environments where external infrastructures are unreliable and maps keep changing. Insuch cases the robot should be able to localize using their on board sensors. IMU sensors are most suitabledue to their cost effectiveness. We propose a novel approach that aims to improve the accuracy of IMU basedrobotic localization by analyzing the performance of gyroscope and encoders under different scenarios, andintegrating them by exploiting their advantages. In addition the angle computed by robots to avoid obstaclesas they navigate, is used as an additional source of orientation estimate and appropriately integrated using acomplementary filter. Our experiments that evaluated the robot over different trajectories demonstrated thatour approach improves the accuracy of localization over applicable existing techniques.