Review of Reinforcement Learning in Nuclear Imaging and radiation sensing waste using Unmanned Autonomous System Navigation

Abstract

In a world where the immense potential of nuclear energy is accompanied by the inherent risks of radiation release, diligent and unwavering monitoring of nuclear radiation ensures a secure future for generations to come. Nowadays, radioactive sources are not just limited to well-maintained facilities, as they are newfound in places such as medical facilities, military, and research laboratories. Moreover, such radioactive materials can also be lost during transportation which poses a risk to human welfare. Therefore, it is important to intercept such radiation sources to mitigate a major catastrophe. With the advent of Industry 4.0, nuclear operators are actively exploring technological advancements to address the challenge of detecting, localizing, and tracking these radiation sources by replacing labor-intensive manual methods that can be expensive and threaten the safety risks of human operators. The use of unmanned autonomous systems (UAS), including unmanned aerial vehicles (UAVs) or unmanned ground vehicles (UGVs), has the potential to provide a cost-effective, accurate, and safe approach to intercept or localize a radiation source.