Research On The Key Technology Of Attitude Control Of Search And Rescue Underwater Vehicle

With the further exploration and development of the ocean,there are more and more accidents in the ocean.Therefore,it is very important to build a complete Marine search and rescue system.In this thesis,based on the independently developed search and rescue Vehicle(ROV)and combined with the actual search and rescue mission,we deeply analyzed the influence of ROV speed on system parameters,developed a search and rescue ROV experimental system,and studied ROV attitude solution algorithm and attitude control strategy.It can improve the attitude control precision of ROV during high-speed navigation and fixed-point operation,so that ROV can be better applied in search and rescue field.Firstly,the background of search-and-rescue ROV is fully investigated,and the detailed search and rescue process of ROV is analyzed.Combined with the research status at home and abroad,the control difficulties of ROV in underwater search and rescue missions are introduced,that is,the attitude control accuracy of high-speed navigation and fixed-point operation is poor.The deficiencies of traditional control algorithms are analyzed,and a more advanced ROV attitude control method is proposed.Secondly,force analysis was carried out on the search-and-rescue ROV,and a six-degree-of-freedom kinematics model was established according to Newton’s law and the law of moment of momentum.The influence of ROV speed on system parameters is analyzed,and the simulation framework of attitude control is given,which lays a foundation for the design of the attitude controller of the AUV.Again,for single sensor for attitude algorithm accuracy is poor,the lack of stability problems,this thesis USES the complementary filter and gradient descent algorithm for multisensory data complementary and optimize the combination of information processing,and aiming at the problem of slow convergence speed of traditional gradient descent algorithm,is put forward based on variable step size momentum gradient descent algorithm,The accuracy of the improved algorithm is compared with the actual test of attitude data processing board.Then,the SAR ROV attitude control based on reinforcement learning algorithm is studied,and the attitude controller based on DDPG algorithm is designed.Aiming at the problems of low model efficiency and low sample utilization rate after the iteration of DDPG algorithm,an improved DDPG algorithm is proposed by introducing noise and variable capacity experience pool in parameter space.By comparing the traditional DDPG algorithm and the improved DDPG algorithm,the simulation experiment proves that the improved DDPG algorithm is superior to the traditional DDPG algorithm and can be applied to the SAR ROV attitude control.Finally,search and rescue the ROV control system was developed,mainly including power box,umbilical cord cable,main control board,from the control panel,posture data processing board,the ROV control system designed for points module debugging and alignment of the land,after debugging at the school pool,lake and Hankou waterfront underwater experiments,the ROV is good for accomplish the action and underwater observation function,Achieve the expected standard.