A Review on Localization Algorithms of Mobile Robot in Different Environments

Abstract

The mobile robots' localization algorithms are considered the main part of robots to make them self-driven. However, most of the localization algorithms have a common problem related to the noisy reading of the sensors during Simultaneous Localization and Mapping (SLAM) of the environments. The noise produces errors in the estimated path which will lead to wrong decisions when handling the localization process. Thus, there is a need for an algorithm to eliminate the noise and correctly estimate the robot's position along with the movement. This paper classifies localization algorithms into three groups, namely, Kalman Filter based approaches, Statistical based approaches, and Artificial intelligence-based approaches. The reviewed algorithms have been arranged from the oldest to the newest with their results, researchers' methods to treat the noises, and tools for sensing the environment (camera, IMU, LiDAR, LRF, and Ultrasonic). One can notice that Kalman Filter-based approaches were used rarely in the previous years, while the statistical-based approaches were combined with other calculations to enhance their performance. The modern approaches used nowadays are AI-based approaches, especially fuzzy logic algorithms. The results of the reviewed algorithms proved that the noises are still affecting the algorithm's performance even though one can use modern algorithms to eliminate all the noisy readings. The laser simulator logic algorithm that has been used in the Active Force Control (AFC)  of mobile robots gave the best results in eliminating the noises. This work reviews most localization algorithms and classifies them based on how they are used for pose estimation; it can be a useful reference for researchers who will work in this field in the future.