Answers are grounded in this paper and cite their sources.
Search and rescue (SaR) missions often use unmanned aerial vehicles (UAVs) with cameras for finding targets in unknown cluttered environments. Due to the unpredictable nature of these environments, reinforcement learning techniques can be used to perform these tasks. In this paper, we employ reinforcement learning algorithm integrated with a position controller to locate a missing victim in a SaR scenario using UAV. To deal with large state space sizes, we use function approximation to achieve faster convergence. The results show how the UAV can successfully carry out the task in a realistic urban SaR scenario. However, common approaches limit their work to target search. In practice, searching for mobile targets is not enough as their position may change before the rescue team arrives for evacuation. To address this, we uniquely combine YOLO with Optical flow to allow the UAV to track and follow the target victim. Extensive simulations demonstrate its application in SaR missions.
No takes yet. Share an insight, caveat, or question.
Ajmera et al. (2020) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: