Research On Path Planning Of Unmanned Aerial Vehicle In Complex Environment

At present,UAV system is attracting more and more attention from many circles.Many scholars and researchers have studied all aspects of UAV technology,and the UAV industry has seen an unprecedented development.Path planning and obstacle avoidance technology for unmanned aerial vehicle are important components to achieve autonomous flight.UAV’s path planning is a safe flight path which satisfies the constraints conditions according to the goal of task,which is the key point to ensure the successful execution of the task.Currently,the track planning system is to design a safe flight track from the starting point to the ending point for a single unmanned aerial vehicle.While for multiple UAVs path planning,the system is not only to consider the physical properties of a single UAV,but also to consider the number of UAV and the assignment of the task.The key technology of UAV obstacle avoidance system is the perception of obstacle depth.The depth information needs security assessment and further obstacle avoidance strategies.Therefore,according to the above considerations,the main works are as follow:Firstly,a cost model of path planning for multi UAVs is established according to the analyzes of path planning space modeling method and the constraint conditions for path planning of multi UAVs,such as distance,flight altitude,flight time,task execute efficiency,environmental landscape,etc.Second,a swarm intelligence algorithm for path planning of multi UAVs has been researched.Considering for the shortcomings of the basic particle swarm algorithm,adopting the strategy of adaptive adjustment of inertia weight and hybrid the differential evolution algorithm,an improved hybrid particle swarm optimization algorithm is given,which is compliant the multiple UAVs path planning.Then,the collision detection algorithm about the obstacles between the track nodes has been discussed,building the middle track node of obstacles’ expanding bounding box and external tangential circle of hemispherical threat area to carry out the obstacle avoidance flight.The situation of same or different obstacles between the track nodes has been analyzed.At last,according to the simulation and comparison from multi UAVs path planning algorithm,the most suitable paths have been selected by simulation results.Finally,the depth of obstacle is explored by choosing binocular camera sensor.The relationship between the various coordinate systems and the principle of binocular calibration as well as distortion correction is analyzed.The accuracy of binocular correction is guaranteed by introducing the constraint of polar line.The binocular problem matching and the method of depth map transformation are analyzed.On this basis,the depth information of target point is obtained.In addition,the UAV binocular visual system is designed,which is divided into binocular vision module and flight control module.Through serial communication the information of depth of obstacle is transferred,and the flight experiment for UAV is completed.