Autonomous Path Planning Of UAV For Indoor Unknown Environment

With the advent of the era of intelligence,safe and effective indoor post-disaster rescue,fast and convenient indoor contact-free distribution,accurate and efficient indoor panoramic map construction and other scenes put forward more practical requirements for the application of UAV in the indoor field.UAV path planning is the basis of all indoor operations,so it has important scientific research and practical significance to carry out UAV path planning research in indoor unknown environment.At present,there are still many serious challenges in UAV path planning research,especially in the indoor scene with narrow space and dense obstacles,which poses a greater challenge to UAV autonomous navigation.Pathfinding algorithm shortage caused by local heavy planning is unsustainable and real-time visual trajectory scenarios obstacles caused by unpredictable dynamic obstacle avoidance ability is insufficient,and the optimal trajectory is difficult to get a smooth flight in the UAV kinetics and environmental constraints,these are the indoor UAV route planning needs to solve an important problem,according to the above problem,In this paper,the indoor path planning of UAV is studied mainly from two aspects of front-end discrete path finding algorithm and back-end continuous trajectory optimization.The specific research content is as follows:1)on the characteristics of indoor UAV static global discrete track points exist in the rough,three-dimensional space search time is too long is bad for real-time programming problem,this paper proposes A node based on arc distance weighting adaptive improved A*algorithm,through the use of more accord with the actual trajectory of UAV arc distance as A heuristic function cost evaluation value,The actual value of UAV is more closely coordinated,so as to effectively reduce the expansion nodes and reduce the search time.In addition,according to node location information,this paper adaptively adjusts the weight ratio of actual cost and predicted cost in the evaluation function,so as to allocate the weight of each stage function more reasonably.Experimental results show that the actual search efficiency of the improved A*algorithm is more than 30%higher than the traditional algorithm,and the planned path length is about 2%higher.2)Aiming at the problems of difficult prediction of dynamic obstacles and low success rate of obstacle avoidance in the dynamic global planning of indoor UAV,this paper proposes a local dynamic path planning algorithm based on trajectory prediction model.Firstly,three kinds of modeling classification are carried out according to the actual movement characteristics of indoor obstacles so as to cover more types of obstacles.And established the interaction based on plant-based,multiinput IMM obstacle trajectory accurately predicting the kalman filter model,on the basis of this puts forward a method of dynamic obstacle avoidance,can solve the problem of UAV local dynamic obstacle avoidance,the simulation results show that the proposed algorithm of dynamic obstacle avoidance effectively the success rate reached more than 70%,compared with traditional obstacle avoidance algorithm have made great progress.3)Aiming at the problems of trajectory crossing and unreasonable time allocation in indoor UAV trajectory optimization,this paper proposes a time allocation algorithm based on velocity adaptive adjustment.Firstly,a dynamics model of UAV is established based on the differential flatness of quadrotor system.On this basis,the incompleteness of collision free trajectory optimized by minimum Snap polynomial is demonstrated.Then introduce the minimum Jerk trajectory optimization algorithm based on bezier curve,and on the basis of the bezier control point to design a series of effective constraint conditions,the paper analyzes the time allocation for optimal trajectory and the influence of dynamic index,based on the relationship between changes in the velocity of the node in the span of time distribution for iterative adjustment,from the optimal time distribution.Simulation results show that the improved trajectory optimization algorithm in this paper is more stable than other algorithms,and the planning time and planning path have made corresponding progress.Finally,this paper carries out system simulation verification on Gazebo simulation platform,and carries out UAV path planning flight test in A real indoor scene.Both system simulation and actual flight test show that the improved A*algorithm achieves better search efficiency and greatly shortens the front-end path finding time.The dynamic obstacle avoidance strategy based on trajectory prediction can avoid dynamic obstacles more reliably.Meanwhile,the improved back-end optimized trajectory is more stable and safe and conforms to the UAV dynamics characteristics,which accelerates the landing of UAV indoor application to a certain extent.