A novel augmented reality and reinforcement learning empowered communication framework for underwater unmanned autonomous vehicle

These days, autonomous uncrewed underwater vehicles (UUVs) play a crucial role in marine exploration, surveillance, and environmental monitoring. However, their communication and object identification are key challenges due to high latency, limited bandwidth, and security vulnerabilities. Traditional UUV frameworks have distinct limitations and pose challenges for dynamic communication in critical environments. To address the above issue, this paper presents a novel augmented reality and reinforcement learning-enabled communication framework for UUAV applications to improve communication quality, enhance object detection, and identify system vulnerabilities. In this framework, we propose adaptive augmented reality and reinforcement learning scheduling strategies (AARLSS) to optimize communication at long and short ranges during navigation and to identify objects and vulnerabilities at runtime while executing applications. AARLSS optimises the performance of UUAV, minimises energy consumption and delay, reduces security risks, and improves the accuracy of objective detection. AARLSS offers various methods and functionalities, including using other sensors as inputs, preprocessing, and training the entire workload as a mini-benchmark using deep Q-learning (DQN). A scheduler allocates them to available resources before execution, subject to time and deadline constraints, and verifies them using an adaptive intrusion detection system (IDS). We created an augmented and virtual reality testbed for the experimental setup and evaluated the performance of different methods. The results show that the proposed methods minimised UUAs' energy consumption by 20 to 21%, reduced delay by 18 to 20%, and improved accuracy by 97 to 98% during experiments on the testbed setup.