Research On Adaptive Control Method Of Operational Autonomous Underwater Vehicle Manipulator System Based On Online Approximation Strategy

With the increasing consumption of non-renewable resources,the development of marine resources has become the focus of attention of all countries.The development of marine resources requires advanced technology and equipment,and underwater robots play important roles as the only system that can be operate in deep-sea environment.And the Autonomous Underwater Vehicle(AUV)which has the characteristics of large range of motion and flexibility has become important equipment for marine resource detection.Generally,it does not have underwater operating ability.However,with the acceleration of the marine development process,the AUV equipped with a manipulator has become one of the important research directions in the field of underwater robots,and it also has important practical value for the research of AUV operation technology.This dissertation will investigate the motion control of AUV,manipulator control and the attitude stability control of AUV under the underwater operation state,the detailed research works are presented as follows.Study on the experimental prototype of the operational AUV.Firstly,a joint drive mechanism of "linear cylinder+three-link+four-link" is proposed to solve the long arm extension and folding storage of the manipulator.Secondly,for the problem of the large attitude change of AUV caused by underwater manipulator operation,the attitude control device based on the change of the gravity center is studied to realize the attitude stability control of micro-miniature AUV.Finally,the operation performance of the "UVIC-I" was verified by the pool experiment,such as the deployment and storage of manipulator,the attitude adjustment and the underwater shearing.Study on the water hydraulic system of the underwater long-arm manipulator and its accuracy measurement.Due to the long-arm and the large number of degrees of freedom of the manipulator,the slight error of the joint angle will have a great impact on the end accuracy.To solve this problem,a "closed circulation" hydraulic system is developed,which has a small design flow rate and reduces medium leakage through the medium "closed cycle" and internal dynamic sealing structure.And the water medium is used to reduce the pressure loss of the hydraulic system.In addition,in view of the joint delay caused by large compressibility of water medium,a "vacuum+pre-pressurization" water medium treatment method is proposed.Aiming at the end precision measurement of the manipulator,the author proposes a contact-type end precision measurement method for underwater long-arm manipulator.Finally,the effectiveness of the proposed method is verified based on the pool experiments of"UVIC-I".Study on the adaptive control method of the underwater long-arm manipulator for the operational AUV.For the system model uncertainty,external disturbances and time-delay of joint,which affect the control accuracy of the manipulator,the typical robust control method based on adaptive function approximation has large steady-state error and motion lag when the joint control of the manipulator is conducted.To solve this problem,an adaptive sliding mode PID control method based on Legendre polynomial function approximation is proposed for the underwater long-arm manipulator.In this method,the Legendre polynomial function is used to approximate the uncertainty factor on-line,and derives the joint controller based on the PID sliding mode surface with integral and differential functions of the tracking errors,to reduce motion lag and tracking error.For the function approximation error compensation,the upper bound of the approximation error is adjusted online by the upper bound error control law to ensure that the joint error is consistently asymptotically zero.Finally,the effectiveness of the proposed method is verified by the pool comparison experiment of the underwater long-arm manipulator.Study on the trajectory tracking control method of operational AUV.Aiming at the problems of AUV trajectory tracking control under the influence of ocean current disturbance,dynamic modeling uncertainty,and thrust model errors,the typical sliding mode control method based on forward neural network has long convergence time and large overshoot for abrupt trajectory tracking.And also,the typical method does not consider the effect of thrust model error on trajectory tracking.To solve the above problems,an adaptive backstepping terminal sliding mode control based on recurrent neural networks is proposed.In this method,the regression neural network is used to classify and approximate the ocean current disturbance,dynamic model uncertainty and thrust model errors,and to design the trajectory tracking controller based on non-singularity backstepping terminal sliding surface.At last,based on stability analysis,an adaptive control laws are obtained for on-line adjustment the weights and the scale factor.For the chattering problem of the sliding mode control,a chattering-reduction method of control variables is proposed,which realizes the continuity of sliding mode switching function through sigmoid function,and the sliding mode switch gain is adjusted on-line through the exponential function of sliding surface.The effectiveness of the proposed method is verified by simulation experiments and the pool experiments of"UVIC-I" experimental prototype.Study on the adaptive region control method of the attitude control of the operational AUV under the underwater operation.For the attitude control of the AUV under the influence of current disturbance,system model uncertainty and disturbance of manipulator operation,under the control of the typical attitude adaptive region control method,the attitude angle fluctuates at the boundary for a long time,which affects the attitude control accuracy,prolongs the adjustment time,and increases the energy consumption.Aiming at this problem.an adaptive sliding mode dynamic region control method for AUV is proposed.In this method,a dynamic target region without region boundary is designed,and then combines the dynamic target region with the sliding mode control surface.Based on RBF neural network,the uncertain unknown terms of AUV system model are approximation estimated on-line.Meanwhile,an adaptive adjustment method is proposed to obtain the network weight parameters,radial basis function center and variance on-line.Finally,based on the sliding mode control term,the approximation error is compensated to ensure that the uniform asymptotic convergence to zero.The effectiveness of the proposed method is verified by the attitude control comparison experiment of "UVIC-I".