Research On Joint Communication And Positioning Resource Scheduling Method For UAV-Assisted Emergency Command Sites

During emergency rescue,weak infrastructure and complex terrain impede rescue capabilities.Unmanned Aerial Vehicles(UAVs)have immense potential in providing communication and positioning services to rescuers due to their Agility.However,emergency rescue exists several challenges.First,there is the issue of perceiving disaster sites and demand where communication and positioning demands vary between different disaster areas.The challenge lies in accurately dividing the communication and positioning demands of rescuers in different areas using limited information.Second,there is the issue that rescuers have concurrent and different communication and positioning demands.The question then becomes how to effectively utilize limited bandwidth resources.Therefore,this thesis mainly focuses on the mentioned problems.Firstly,we propose a disaster awareness situation map construction and communication-positioning requirements classification method to address varied demands in disaster areas.Specifically,we employ UAVs to efficiently perceive real-time maps of disaster areas with high freshness.Due to large rescue area and changing disaster conditions,we utilize multiple UAVs to capture images and improve perception efficiency.These images are then stitched together based on coordinate information to detect the entire disaster situation on the map.Subsequently,we use a machine learning model trained on detection results,classification criteria and expert knowledge to classify communication and positioning needs of rescuer in different areas.By combining expert knowledge with intelligent machine learning algorithms,our model considers multiple factors that affect emergency communication and positioning.Additionally,our model supports fine-tune based on situational changes to further performance.This assists command in better understanding the overall situation while rationally allocating resources based on the extent of disasters in different areas.Then,we propose a joint communication-positioning adaptive method for UAV networks to address the concurrent and time-varying demands of rescuers.Our solution utilizes scarce spectrum resources to meet the service demands.We establish a utility function that integrates communication rate and positioning error while jointly considering single coverage constraints for communication and triple coverage constraints for positioning.Based on this,we propose a genetic algorithm-based multi-agent deep deterministic policy gradient method to adapt UAV deployment,role switching and user link-building strategies.Experimental results demonstrate that our method effectively improves communication-positioning utility compared to other methods.Finally,to validate the effectiveness of the communication and positioning needs classification theory and the UAV network adaptation scheme,we constructed a UAV network dynamic deployment system for verification.Specifically,the system adopts a browser-server separation architecture,where the browser provides UAV parameter and user parameter configuration,and the server is deployed based on the Flask framework.Based on the parameter information passed from the browser,we use the algorithm to visualize UAV position deployment,role switching,and user-link control.This can effectively serve the emergency command site and improve emergency rescue efficiency.