Position Estimation for Mobile Robots Using Machine Learning

Abstract

Robot localization is a critical challenge present in autonomous robots to estimate their position within the environment, ensuring accurate navigation and task execution. This research project compares the position estimation capabilities of advanced machine learning methods Gradient Boosting and K-Nearest Neighbors. In this work Machine Learning (ML) models are built using a dataset from zenodo.org that tracked Received Signal Strength Indicator (RSSI) readings and tested them against strong performance measurements. This work findings show GB provides better results than KNN because it reaches an R2 score of 0.9341. These insights are applicable in various fields, including autonomous vehicles, industrial automation, service robotics, and multi-robot coordination, guiding the selection of appropriate algorithms for specific localization challenges.