The Research On Global Path Planning Of Autonomous Underwater Vehicle

An important research field of Autonomous Underwater Vehicle(AUV) is global path planning(GPP) which is to find a collision-free path from the source position to the destination position among known obstacles. Genetic algorithm (GA) is such an algorithm that simulating the principle of biology evolution. As far as we know, the convergence of standard genetic algorithm (SGA) tends to get into a local optimum. Aiming at this disadvantage, modified genetic algorithm(MGA) adds three operations—restoration, reconstruction and recording the better to the SGA to make the algorithm converge to a global optimum without the change of the search randomicity. The main content of this paper is concerned about solving the GPP problem of AUV with MGA.In this paper, the current situation of AUV's development and the background and significance of GA is introduced, and all the methods in solving GA problems are discussed as well. The author also introduces the basic principle, conceptions, characteristics and other contents of GA. Then, based on SGA method, MGA method was introduced .It was applied in the 2-Dimensions GPP problem and in the 3-Dimensions GPP problem. Thereinto, 2-Dimensions environment is constructed by grid environmental model, and the genetic operator is modified ; 3-Dimensions environment is constructed by hierarchical model, and the probability of the genetic operator is computed by self-adapting probability formula. Finally, MGA method was tested by the joint simulation experiment in the AUV simulation system, the validity was proved by the result of simulation, at the same time, the performance of the algorithm was analysized, the results of SGA and MGA were compared in the same environment and condition, it was proved that MGA is better than SGA.