Application of machine learning for autonomous robots in a simplified environment

Abstract

Robomaster University AI Challenge (RMUA) is an annual competition co-hosted by DJI, IEEE, and the International Conference on Robotics and Automation (ICRA). The most recent advancements in AI are implemented and highlighted in this competition. One area that where AI can be used to improve upon the existing technology is localization. This paper aims to use machine learning as a method of sensor fusion to localize a robot. Furthermore, in this project, the machine-learning based localization method will be benchmarked and implemented directly with the navigation system. The result shows that the implementation of convolutional neural network as a sensor fusion method shows promise of improving the existing localization methods.