Research On UAV Path Planning Technology Based On Cloud-based Adaptive Genetic Algorithm

Genetic algorithm is a kind of randomly optimized search algorithm which is on the basis of biological evolutionary theory in the nature, and it can solve different kinds of function optimization problems. The advantages of this algorithm lie in: first of all, encoding technology and genetic operation are relatively simple,and restrictive conditions are less;secondly, the parallel and global search ability of the algorithm is relatively strong. However the traditional genetic algorithm has many shortcomings, for example, the crossover operator and mutation operator of the algorithm are strongly random and blind, thus this algorithm is prone to premature convergence, slow convergence speed, poor local search ability and so onThe adaptive genetic algorithm is proposed based on these above problems, the existing adaptive genetic algorithms can be general divided into two categories: linear adaptive genetic algorithm and nonlinear adaptive genetic algorithm. Linear function is used to self-adaptive adjust the crossover probability and mutation probability by linear adaptive genetic algorithm,it is very disadvantageous to increase the population diversity and keep the good individual,even the algorithm converge slowly, and it is more likely to get the local optimal solution.Nonlinear function is used to self-adaptive adjust the crossover probability and mutation probability by nonlinear adaptive genetic algorithm, this kind of algorithms do not have above disadvantages of linear adaptive genetic algorithm, but with the same linear adaptive genetic algorithm, the parameters of these algorithms need to be set according to many experiments,which is not conducive to the operation of complex and volatile environment. In addition,both linear and nonlinear adaptive genetic algorithms have a common weakness, it is that a determined individual fitness corresponds to a unique crossover probability and mutation probability, and it is not conducive to jump out of local optimal solution.However cloud theory is a kind of uncertain conversion which is between qualitative knowledge description and quantitative data expression, and it is very suitable for dealing with uncertain problems,so the theory of cloud is introduced into the adaptive genetic algorithm in this paper,and the X-condition cloud generator is used to improve the adjustment formula of the probability of crossover and mutation, so an adaptive genetic algorithm which is named cloud-based adaptive genetic algorithm is constructed in this paper. Because of randomness and tendency of cloud droplets in the cloud model, the crossover probability and mutation probability both have the tendency of the traditional genetic algorithm which can satisfy the strong searching ability and a certain degree of randomness. When the individual fitness is a certain value,the probability of crossover and mutation are not fixed data but random values which follow a specific distribution. This can not only avoid algorithm falling into local optimal solution, but also can improve the diversity of the population.The cloud-based adaptive genetic algorithm is applied to UAV path planning in this paper, and the result of path planning of UAV under incomplete known environment come out,the result of simulation demonstrates that the cloud-based adaptive genetic algorithm is more effective than the existing adaptive genetic algorithm in the application of UAV pathplanning.