Research On Visual Based Robotic Hoisting Motion Following Control

The marine economy has played a very important and positive role in the national economy and social development.As China's investment in marine logistics and projects increasing,the lifting equipment of marine logistics and projects is also developing rapidly.Lifting equipment is one of the main equipment which is widely used in marine logistics and projects.The safety and efficiency of lifting and lowering will directly affect the operational efficiency and economic efficiency of the logistics and projects,and the complex oceanic sea state is one of the important influencing factors.Therefore,ensuring the safety and efficiency of cargo transfer and lifting equipment is the research direction and research hotspot in marine logistics and engineering.In order to improve handling efficiency and safety,building a comprehensive experimental platform for marine logistics equipment intelligent lifting can be used for studying on lifting equipment control,marine or port intelligent lifting or transfer technology and lifting equipment performance evaluation and optimization in the laboratory environment.In this paper,based on the comprehensive experimental platform of marine automatic and intelligent lifting equipment which is built by the relevant research group of Jimei University and the research group of Jilin University,the visual-based hoisting robot arm winch lifting and following control method is studied.The main work of the paper is as follows:1.System composition and design.According to the laboratory platform and functional requirements,the hoisting mechanism on the hoisting robot arm is designed,and the visual sensor is integrated into the hoisting mechanism,the overall construction of the hardware and software platforms required by this paper is achieved.2.Kinematic analysis and hand-eye calibration.Based on the lifting machine arm,the kinematics and the inverse kinematics is analyzed by using the D-H parameter method and conformal geometry algebra method with a case for validation,respectively.The restricted work space by field of view is analyzed based on Monte Carlo algorithm.First-order distortion correction model was selected for camera calibration,and ParkMartin algorithm was used to calibrate the ‘hand-eye' relationship between robot arm and winch.3.Visual marker detection and localization.By using visual marker the target objects in complex environments can be detected and identified.The main direction and direct linear transfer algorithm is used to calculate the orientation and position of the marker.A combination marker resisting image noise and occlusion issues with pose confusion scheme is designed.A quaternion based Kalman Filter is used for pose estimation of combined marker.4.Winch following control strategy.Based on the double S-curve motion planning algorithm and its acceleration/deceleration method,the detected target motion is tracked and preplanned.With establishing winch model,the pseudo-derivative feedback and P-feedback compound control are used on the velocity and position loop,respectively.The overall winch following strategy is described.5.Visual-based winch lifting and following control is implemented based on Visual Studio 2015 and Qt5.8.Based on the above-mentioned chapters' work,the proposed winch control strategy was validated and the experimental results were analyzed by designing static targets and dynamic targets following experiments.