Design Of Underwater Weightlessness Simulation Manipulator And Its Visual Positioning Technology

This paper is based on the preliminary research project of the scientific research project "Development of underwater manipulator" of Beijing Astronaut Research and Training Center.This manipulator is used for underwater weightlessness simulation test and training to support astronauts' extravehicular activities.In this paper,the overall scheme of underwater weightlessness simulation manipulator(hereinafter referred to as underwater manipulator)is designed,and the structure of underwater manipulator is designed and verified,and the key technology of visual positioning of underwater manipulator is deeply studied.Firstly,this paper introduces the overall composition,control system and software monitoring interface of underwater manipulator,and focuses on the design of visual control system,including the construction of hardware and software platform,the configuration mode of camera and manipulator,the layout of markers and two design schemes of visual control system.It provides experimental platform and research for the follow-up structural design and visual positioning research.Direction.Secondly,the structure design of underwater manipulator is carried out.The key technologies are underwater dynamic seal drive and neutral buoyancy balance.Firstly,the necessity and feasibility of the driving scheme of magnetic coupling seals are analyzed,and the underwater magnetic coupling seals are analyzed and designed in detail.Then the force analysis of the underwater manipulator is carried out to determine the key dimensions and the selection of key components of the manipulator,and the structural design of each joint and base of the underwater manipulator is completed.On the basis of these structural designs and parts selection,the neutral buoyancy balancing design of underwater manipulator is carried out by using the modeling software Solidworks.Finally,the rationality of the structure design of the underwater manipulator is verified by the joint simulation analysis of Adams and Simulink,that is,the structure design of the underwater manipulator meets all the relevant indicators in the task book.Thirdly,since the joint of the underwater manipulator is driven by flexible magnetic coupling,and the underwater manipulator itself has elasticity,and elastic deformation will occur under the action of water resistance,it is necessary to implement closed-loop visual servo control to achieve the required positioning accuracy of the underwater manipulator's end.In this paper,the key technologies of vision positioning of underwater manipulator are studied,including underwater image enhancement,image segmentation and target recognition.The image captured by the camera on the underwater manipulator has the characteristics of uneven illumination,low contrast and random noise.According to these characteristics,image processing is carried out separately.Homomorphic filtering improves the uneven illumination of the image,histogram stretching enlarges the contrast of the image,and median filtering effectively eliminates noise and preserves the edge of the object.The experimental results show that the image segmentation method in this paper is superior to other segmentation methods.An improved RED ellipse detection algorithm for underwater environment is proposed.According to the analysis of experimental results,the processing speed and accuracy of ellipse detection can meet the key requirements of real-time and stability of the control system.Finally,the positioning accuracy is improved by establishing the underwater refraction imaging model of the camera and improving the calibration method of the underwater camera.Because the imaging model of underwater camera is different from that of land camera,it is necessary to consider the distortion caused by refraction.According to Snell's law,the distortion caused by refraction is classified into higher order distortion coefficients,which effectively completes the establishment of underwater camera imaging model.After that,an improved Zhang's calibration method suitable for underwater environment is proposed when calibrating the distortion coefficient of the camera.This method solves the difficulty of low accuracy of initial iteration value of Zhang's calibration method.The experimental results show that the calibration accuracy of this method is higher than that of traditional Zhang's calibration method.After considering the imaging model of underwater camera and the improved camera calibration method,the positioning accuracy of elliptical target is measured.The results show that the repetitive positioning accuracy of the camera can reach 4.24 mm,which meets the key requirements of the positioning accuracy of the underwater manipulator.