Unmanned Aerial Vehicle Route Planning Assessment

As one of the critical components of Mission Planning System,Route Planning plays a key role in realizing intelligent navigation for unmanned aerial vehicles(UAV).As the flying environments very complicated,route planning is particularly important to improve the survival rate of UAV,making sure the mission is successfully completed.Route planning involves many factors that are coupled with each other,so the planned route still needs to be assessed.Similarily,route evaluation involves a number of factors.All related factors are taken into account to deliver comprehensive,objective and fair evaluation.Route planning algorithms known in the field include dynamic programming algorithm andintelligent optimization algorithm.Dynamic programming algorithm will take up too much space,while intelligent optimization algorithm has uncertain convergence time and other shortcomings.Route evaluation algorithms include Simple Additive Weighting Method(SAW),Analytic Hierarchy Process(AHP)and so on.They all revolve around the judgement matrix.This paper improves adaptability of route planning by the current algorithm,and evaluates routes through FCE,BP neural network and BP neural network that optimized by GA.The main works of this paper are as follows:Firstly,the paper establishes an arithmetic model based on route planning,makes an analysis from the aspects of space model and restricted conditions,and plans for route by the three-dimensional sparse A * algorithm.The simulation analysis to the simulation experiment of route planning after changing conditions shows that routes varies under different planning alternatives.Through the route planning simulation experiment,some important factors influencing routes can be obtained and used as important index for route evaluation.Secondly,There are many factors influencing the route evaluation.The paper at first analyzes the factors influencing the route to clarify hierarchical relations of indexes,and establishes an index system of UAV route evaluation with hierarchical structure.Because of different units and orders of magnitude of each index,we need to normalize indexes.The paper presents the normalized model of indexes,weighing over the relative importance of indexes by AHP and combining FCE to evaluate merits of the route.The simulation analysis shows that objective and fair evaluation of the route can be made through FCE.Thirdly,based on the model of weapon effectiveness evaluation,the paper applies BP neural network to route evaluation.Through analysing the performance of each training algorithm is by simulation experiment,LM training method is used to train the network.The simulation proves that correct evaluation of route can be basically obtained by the BP neural network,but the results have error and differ every time with instability.In view of shortcomings of the above evaluation algorithm,the paper adopts the genetic algorithm to optimize the initial weights and initial thresholds of BP networks to obtain a network with better performance.The simulation of two different conditions shows that the optimized BP neural network can correctly evaluate the route of the UAV.The genetic algorithm works well in optimization and leaves a fine evaluation effect.,Algorithm-based evaluation is stable.