Enhancing Urban Unmanned Ground Vehicle Operations Through Machine Learning

This paper investigates the use of reinforcement learning (RL) to improve how unmanned ground vehicles (UGVs) explore different environments, including urban out-door, indoor, and open terrain. The paper utilizes a simulation setup developed and explored in the previous work by the authors in attempt to assess how RL can enhance exploration strategies in terms of efficiency and adaptability in urban areas. The results show that RL-based methods improve UGVs' ability to navigate and survey urban environments compared to traditional approaches. This improvement underscores the potential of RL to advance the capabilities of autonomous systems in complex settings. While the current focus is on exploration, future work will look into using RL for more complex scenarios involving direct engagements, such as encounters with enemy forces. The goal is to develop more sophisticated RL models that can manage multiple objectives, enhancing UGVs' functionality for urban operations.