Research On UAV Formation Obstacle Avoidance Algorithm For Urban Post-disaster Rescue

Nowadays,as the process of urbanization continues to accelerate,more and more people are entering urban life.In order to accommodate more people,there are more and more buildings in the city.However,China has a vast territory.Among them,the changeable topography and complex geological structure have caused frequent geological disasters in our country and seriously threatened the lives and property safety of the people.Once a large-scale geological disaster occurs in a city,emergency rescue is required.The high-rise buildings in the city will inevitably block the view of the large-scale drone at high altitudes,and it is impossible to detect the specific disaster situation and find the trapped persons.It is necessary for small UAVs to fly at low altitudes in post-disaster urban scenes.However,a single UAV is often difficult to perform complex tasks due to its limited load and flight time.Therefore,UAV formation flying can be used for emergency rescue after disasters.This work has a certain research value.However,the post-disaster city scene is a very complex scene,if UAV formations want to carry out search and rescue operations in the post-disaster urban environment,they must first ensure the flight safety of the UAV formation itself.Therefore,the related research of UAV formation avoidance flight is of great significance in the field of emergency rescue after urban disasters.However,there are few suitable centralized control UAV formation obstacle avoidance algorithms that can be directly used for UAV formation obstacle avoidance and ensure Leader UAV can lead the formation with collision avoidance and fly safely.At the same time,the post-disaster urban scene is relatively complex and extremely unstable.When drones are flying in formation,the formation structure is prone to be destroyed.At this time,Follower drones need to make corresponding decisions to avoid obstacles while maintaining the drone formation geometry.In summary,in order to enable the UAV formation to fly safely and stably in the post-disaster urban scene,the main requirements are as follows:1.A centralized control UAV formation obstacle avoidance algorithm for the Leader UAV,which can ensure the UAV formation to avoid obstacles;2.A distributed control UAV formation obstacle avoidance algorithm is used for Follower UAVs.When the Follower UAV encounters obstacles,it can avoid obstacles while maintaining the formation structure;3.One simulation platform which can support UAV formation simulates in the post-disaster urban scene.Based on this logic,the main research content and work of this article mainly include the following three parts:First,the centralized control of the UAV formation avoidance algorithm.In this paper,based on the VFH algorithm,the geometric constraints of the drone formation are added,and the UAV formation and single UAV’s safe flight minimum circumscribed circle diameter 2R_s,2r_swere as a judgment basis.It is directly compared with the sector aisle’s width,and the sectors whose aisle width is smaller than the circumscribed circle diameter2r_sof the single machine are deleted.When the aisle’s width D is greater than 2R_s,the UAV formation passes directly,and when the aisle’s width 2r_s<D<2R_s,the formation passes through after changed formation.When UAV formations fly in post-disaster urban scenes,the size of the obstacle is often large than the UAV’s vision.At the same time,the VFH algorithm used in this thesis is a local obstacle avoidance algorithm,which has no memory and is prone to oscillate,especially when the size of the obstacle is larger than the field of vision.In response to this problem,this thesis introduces a trajectory correction factor flag on the basis of the VFH algorithm.When the UAV formation selects a safe sector,if the current direction is still safe,then fly forward flag steps otherwise continue the VFH algorithm to select a new safe sector.This method can solve the problem which is easy to oscillate without increasing the time complexity of the algorithm.Second,the distributed control formation obstacle avoidance algorithm.According to the previous analysis,this article needs a distributed control UAV formation obstacle avoidance method that has low time complexity and can maintain the formation geometry structure.Based on this,this thesis simplifies the calculation process of the polar coordinate histogram,which can reduce its time complexity to a certain extent.At the same time,this article introduces a dynamic target.After the drone moves in a safe direction,refresh the new desired position of the Follower drone.If the direction is safe,it will exit the obstacle avoidance program directly,fly to this point,and then enter the formation.Third,build a simulation platform based on the UAV formation after the urban disaster.In order to better test the improved algorithm proposed in the previous article,this paper builds a post-disaster UAV formation simulation platform based on PX4,ROS,and Gazebo.This platform can provide more realistic experimental scenarios for algorithm simulation.At the same time,PX4 and ROS The introduction enhance the versatility of the algorithm.