Automated unmapped forest path navigation of mobile rover using neural networks

Abstract

In this paper, a system for autonomous navigation of unmapped forest paths in a simulation, along with a new simulator for testing and training it is presented. For navigation, it uses a combination of path planning and deep neural networks trained with imitation learning.

Keywords: Navigation, Neural networks

Published on website: 1.9.2022

Attached files: Nikola_Jovicic.pdf