Research On Control Method Of Underwater Manipulator Based On Visual Guidance

The underwater manipulator is an important part of the underwater robot system and plays an indispensable role in deep sea resource exploration and underwater operation.At present,there are still some shortcomings in the underwater manipulator technology: the end of the manipulator is not high in precision and the integration is low.The underwater camera is added at the end of the underwater manipulator to provide feedback information for the underwater manipulator to operate underwater,which can greatly improve the operation precision and self-operating ability of the manipulator.Therefore,the research of underwater manipulator control technology based on vision guidance It has important scientific value and practical engineering significance.Firstly,the DH parametric method modeling and forward and inverse kinematics analysis of the six-degree-of-freedom underwater manipulator are performed,and the Jacobian matrix is solved and the geometric error generated by kinematic modeling is analyzed.According to the actual rotation range of the six rotating joints,The working space of the underwater manipulator is simulated.The dynamics of the underwater manipulator is simulated by Newton-Eulerian iteration method.The control parameter curves of six rotating joints are simulated.According to the actual situation of the underwater,the Morison formula is used for water.The hydrodynamic modeling of the lower manipulator simulates the water resistance torque and the additional mass moment curve of the six rotating joints.Secondly,a kinematics compensation method with parameters is proposed,which solves the disadvantages of the traditional parameter method error,and uses the markers to calibrate the joints of the underwater manipulator.The basic theory of camera imaging is introduced.The distortion is not considered in the process,and a model optimization method based on radial distortion is given.Finally,the basic principle of hand-eye calibration is discussed,and an improved algorithm based on matrix direct product is proposed,which improves the traditional method without considering the system.The disadvantages of noise,calculation error and other factors.Then,based on the theoretical basis of the first two chapters,the hand-eye coordination control system framework and the visual guidance system control flow that mimic the biological vision are proposed.The three-dimensional positioning of the underwater target object is realized by the combination of sonar sensor and monocular vision,and the Condensation algorithm is adopted.The underwater target tracking is realized.The related concepts of auto disturbance rejection control and the corresponding working mechanism are introduced.The auto-disturbance rejection controller based on terminal sliding mode is designed and its stability analysis is carried out.Finally,the actual underwater situation is simulated.The designed controller performs simulation experiments under different conditions such as hydrostatic environment,ocean current environment and external disturbance environment to verify the reliability,stability and robustness of the control method used in this paper.Finally,the overall control system design of the 6-DOF underwater manipulator and the experimental platform of the upper and lower machine are constructed.The F051R8T6 control driver based on STM32 chip is designed.The upper computer based on Qt platform and the three-phase bridge inverter circuit in Linux operating system are designed..The watertight experiment,motion control experiment and the guided robotic arm grabbing experiment based on Unreal Engine 4 platform were carried out by using the experimental platform.The experimental results verified the kinematics modeling of the underwater manipulator.System calibration technology,joint controller design rationality.