Unmanned Aerial Vehicle's Three-dimensional Constrained Multi-objective Trajectory Planning

The widespread use of UAV(Unmanned Aerial Vehicle)has made trajectory planning a hot research area.The UAV trajectory planning is a significant aspect of UAV research field,in the meantime,which is an indivisible part of UAV autonomous system.The favorable trajectory planning methods are able to reduce UAV fuel consumption and improve UAV survivability.Although there have been numerous UAV trajectory planning methods so far,it will be difficult to effectively apply in dealing with trajectory planning problems under complicated conditions on account of oversimplified planning models or inefficient planning methods.In this paper,certain related problems of UAV trajectory planning in 3D static environments are discussed,including trajectory planning model,trajectory planning algorithm,feasible trajectory generation and optimal trajectory generation.The contents and innovations of this thesis are listed as follows:(1)To aim at UAV 3D online feasible trajectory planning problem and UAV 3D offline global optimal trajectory planning problem,combining UAV`s fundamental constraints,diverse mathematical models are established respectively.The feasibility of these models have been validated in the following chapters.When smoothing preliminary flight trajectory,two UAV flight dynamics constraints which based on differential geometry have been put forward reasonably for the sake of making it closer to the actual flight pattern of UAV.(2)To aim at UAV 3D online feasible trajectory planning problem in mountain terrains under threat of wind shear,a 3D online feasible trajectory planning mathematical model originating from part(1)is adopted.When solving aforementioned model by adopting classical RRT-Connect(Rapidly-exploring Random Tree with Connection)algorithm,random node`s generation strategy is single and time consuming,hence 6 kinds of improved strategies for generating random nodes are came up with.Then the improved RRT-Connect algorithms have been applied to the proposed UAV 3D online feasible trajectory planning problem.Compared with using classical RRT algorithm and classical RRT-Connect algorithm to solve the same problem,experiment results show that the improved RRT-Connect algorithm based on the sixth strategy can efficiently generate UAV 3D online feasible trajectory which has shortest trajectory planning mean time.(3)To aim at UAV 3D offline global optimal trajectory planning problem in mountain terrains under interference of communication source signals,a 3D offline global optimal trajectory planning mathematical model stemming from part(1)is introduced.Then the classical multi-objective NSGA-2(Non Dominated SortingGenetic Algorithm-2)and multi-objective ENSGA-2(Efficient Non Dominated Sorting Genetic Algorithm-2)which based on approach to ENS(Efficient Non-dominated Sorting)have been employed in solving the proposed UAV 3D offline global optimal trajectory planning problem in order to verify the rationality and feasibility of proposed model.Compared with using classical multi-objective NSGA-2to solve the same problem,experimental results show that multi-objective ENSGA-2can efficiently generate UAV 3D offline global optimal trajectory which has shorter trajectory planning mean time and more smooth trajectory curve.Meanwhile,experimental results also indicate the rationality and feasibility of proposed model.