Safety Considered Docking Guidance And Trajectory Planning Technology For AUV

The autonomous underwater vehicle(AUV)autonomously performs underwater recovery and docking with the underwater vehicle for recovering,which enhances the physical connection between the AUV and the human occupied vehicle(HOV)in different tasks,and jointly improves the ability both of AUV and the HOV.There are still some problems to be solved in the mobile recovery process of AUV.This paper mainly studies the precise positioning of AUV recycling process,the quantification of recycling safety and trajectory planning.In this paper,the following research is carried out on the above issues:Firstly,a combination navigation system of underwater ultra-short baseline(USBL)positioning system with inertial navigation dead reckoning is designed.A navigation method for manned submersible(HOV)recovery AUV is proposed.Among them,the underwater navigation sensors carried by the AUV include an ultra-short baseline(USBL),a Doppler speed meter(DVL)and an inertial navigation system(INS).By using the information of these sensors,a cooperative navigation module combining USBL positioning and inertial positioning is established.In this study,it is assumed that there is no communication between the AUV and the HOV;therefore,in order for the AUV to obtain the HOV position in the inertial coordinates,an estimation method is proposed.In order to reduce the error accumulation of AUV navigation,a single-step dead reckoning positioning method combining inertial navigation and USBL positioning is proposed.This positioning method is a correction algorithm in data fusion.The navigation accuracy is improved,and the smooth output of the navigation information is ensured.Secondly,the safety is the most concerned issue in the AUV recycling process.For the special environment of the ocean in the recycling process,a method forjudging the safety of multiple factors is proposed,and the safety is quantified.This method of security determination is not only related to the characteristics of the map,but also related to the influence of the current on the AUV and the state of the AUV itself.Thirdly,the recycling process is designed according to the background of the mobile recovery for AUV.The motion of the mother boat is predicted under the premise of precise navigation,and the path planning scene is designed.Because the security decision is related to the state of the AUV,a safety path is planned using the Q-learning method in reinforcement learning.Aiming at the increase of the map range,the "dimensional explosion" problem of Q learning proposes an A*algorithm combined with average security to calculate the sub-sites under the coarse map and generate the connected target areas.In order to find the starting point and the target point in the adjacent target area,the "generalized input-output method" is applied to calculate the feature points of the area.The path point that dynamically reclaims the AUV is finally generated.Finally,after the path planning calculates the path point,the trajectory planning is performed according to the speed constraint and the acceleration constraint of the AUV,and the velocity profile and the acceleration profile of the path are generated.