Research On Global Multi-objective Path Planning And Local Dynamic Obstacle Avoidance For AUV

With the emphasis on marine territories and resource in various countries,various underwater vehicles have developed rapidly.As a new type of underwater detection tool,the autonomous underwater vehicle(AUV)has unique advantages in performing marine tasks because of its advantages such as flexible control,low energy consumption,and high intelligence,and it is widely used in civil engineering and national defense construction.Among the hot topics in AUV related research,path planning technology,as an important part of navigation control technology,is the guarantee of AUV's normal navigation and safe work in a complex and changing marine environment.This paper mainly studies the path planning technology of AUV from both global and local aspects,aiming to realize the multi-objective optimization needs of AUV global path planning and improve the dynamic obstacle avoidance performance of local path planning.Firstly,according to the multi-objective optimization needs of AUV navigation in complex marine environment,a multi-objective global path planning method based on improved ant colony algorithm is proposed.In terms of heuristic function,the traditional heuristic function based on the distance of flight factor is improved to accelerate the algorithm convergence speed,and the energy consumption factor and channel safety factor are increased to form a comprehensive heuristic function;In terms of pheromones,the ant colony pheromone is classified,and the pheromone update mechanism is improved based on the ranking of different target performances to achieve the simultaneous optimization of three goals.And the idea of Pareto solution set is introduced to find the optimal path under different navigation goals,improving the comprehensiveness of the solution;In terms of environmental modeling,in view of the problem of excessive calculation of traditional 3D raster maps,a simplified model of highly dimensionality reduction is proposed,which can retain important information while simplifying the model.Secondly,for the local dynamic obstacle avoidance requirements of AUV in three-dimensional ocean space,a three-dimensional local path planning method based on perturbed fluid model is proposed.Based on the local environment information,the initial flow field model is established,and the correction matrix is used to quantify the disturbance effect of various obstacles in the environment on the initial flow field.Calculate the corrected fluid flow line and use it as the dynamic planning path of AUV.Analyze the main content of the algorithm,and propose a solution to increase the additional correction field for the stagnation point problem that the algorithm may have in practical applications;At the same time,in order to avoid excessive obstacle avoidance and untimely obstacle avoidance of AUV,a rolling optimization strategy is used to optimize the corresponding response coefficient of the algorithm to improve obstacle avoidance efficiency.At the same time,the traditional speed obstacle method is improved.During the path tracking process,the dynamic obstacle avoidance direction guide point is obtained by establishing a three-dimensional collision cone and speed obstacle cone model to guide the AUV to avoid obstacles;After the obstacle avoidance is completed,the parameters of the objective function are adjusted,and the tracking of the reference path is restored,so that the navigation trajectory does not deviate from the reference path due to obstacle avoidance.Finally,for the shortcomings of global optimality and real-time performance of traditional local path planning in terms of dynamic obstacle avoidance,a hybrid obstacle avoidance structure-based AUV dynamic obstacle avoidance planning strategy was introduced.Take the offline global path as a reference,and realize the stable tracking of the global reference path through the designed model predictive controller;At the same time,the traditional speed obstacle method is improved.During the path tracking process,a dynamic obstacle avoidance direction guide point is obtained by establishing a three-dimensional collision cone and speed obstacle cone model to guide the AUV to avoid obstacles;After the obstacle avoidance is completed,the parameters of the objective function are adjusted,and the tracking of the reference path is restored,so that the navigation trajectory does not separated from the reference path due to obstacle avoidance.In the end,AUV navigation can not only make full use of global map information,but also improve the ability of real-time obstacle avoidance.