CBR Study On Visual Decontaminating During Operation Of Underwater Rigid-flexible Hybrid Serial Manipulator

Robots working in an underwater dynamic environment need complex visual perception to adapt to the environment,which requires emergency planning ability to deal with various visual pollution emergencies and improve the reliability of operation.In the above context,the following studies are carried out for the kinematics and visual decontamination-cleaning strategy of the underwater rigid-flexible hybrid series manipulator:Firstly,an emergency strategy was proposed for the visual pollution of the manipulator,which used case-based reasoning.The steps and processes of CBR case-based reasoning applied to the manipulator are analyzed and described.The correct detection and extraction of pollutant characteristic information were realized by combining ViBe algorithm and frame difference method.Then WEKA software was used to create the case library of visual contamination of manipulator,finally the applicable case library was determined by case analysis and comparison of several algorithms.Secondly,the splicing strategy and motion theory of the manipulator’s contaminated field of view are analyzed.The method of field of view stitching is the combination of SIFT algorithm and RANSAC algorithm,and combined with the corresponding movement of the serial manipulator to achieve the field of view stitching effect.Based on the structure characteristics of the underwater rigid-flexible hybrid serial manipulator,the kinematics of the manipulator is analyzed.By combining the piecewise constant curvature hypothesis with the geometric relationship between the manipulator,the forward and inverse kinematics model of the manipulator was constructed,and the mapping between the muscle length and the position and pose of the manipulator end was solved to obtain the position and pose of the manipulator and the workspace.Then the inverse kinematics solution model was established by combining the region representation method and the energy minimization principle,and the results of the inverse solution were verified.Then,the image denoising strategy of case-based reasoning and the cleaning strategy of bionic cleaning device are analyzed and implemented.The improved median filtering algorithm is used in image denoising.The BioTRIZ method was used to map the engineering problems into the biological space,and the bionic visual decontamination device was designed to clean the pollutants.Finally,relevant experiments of various emergency planning strategies and the whole experiment of the series manipulator of CBR after visual contamination were carried out.Carry out pollutant detection in different environments,and verify the feasibility of several schemes based on CBR for different pollution situations.The functions of the thick and thin manipulators and their interaction ability were verified through the series manipulators’ capture of the moving target in water.