Research On Grasping Simulation Technology Of Soft Manipulator Baced On Visual Recognition And Location

Robots have always been a hot field of scientific research.Due to their high flexibility,high adaptability,and high safety,soft robots are highly matched in many aspects of housekeeping services,assistance to the sick and the disabled,and agricultural medical care.Therefore,more and more scholars will study Moved to the field of soft robots.However,the software manipulator does not have a more complete control strategy like the hardware hand.The development of a more complete autonomous control strategy and theory suitable for the software manipulator needs to be solved urgently.It is necessary for the exploration and research of the software manipulator's grasping technology.Machine vision is used in the detection and positioning of targets,industrial products,navigation control robots and other fields.It has the advantages of non-contact,high accuracy,strong reliability,and not depending on the environment.The dexterous hand is a kind of robot end effector.Multi-finger grasping control is a key issue in the research of dexterous hands.How to determine the relative position of the fingertip and the object to be grasped,and how to apply and how much on which position of the object The researcher's analysis is needed for the power and how to judge the grasping stability.The bionic software manipulator designed and developed by this research group has a degree of freedom similar to that of a human hand.It is suitable for life services,agriculture,rehabilitation and health care.It is necessary to ensure that the end effector has high flexibility and safety.Therefore,the research on the software hand is Very necessary and valuable.This topic studies the relevant principles of the camera's three-dimensional positioning,and performs camera calibration and binocular registration on the Kinect color camera and depth camera,which provides a theoretical basis for the software hand grasping experiment and simulation technology.An object recognition algorithm is also studied,and the principle and specific experimental process based on the SSD target recognition method are described.The article also implements the Kinect acquisition,binocular registration and SSD target realization algorithms through the Lab VIEW platform,and uses the visual assistant to recognize Feature extraction is performed on the back frame,which achieves more accurate target positioning.The kinematics and statics of the pneumatic network soft manipulator are studied.It specifically analyzes the kinematics of the soft hand,including the soft hand and its finger workspace,and then analyzes the multi-finger kinematics of the soft hand from the aspects of object,palm,fingers and contact kinematics,and establishes the software hand multi-finger grasping system The kinematic relationship between the grasped object and the finger joints;this paper also analyzes the gripping force closure of the soft hand,selects the soft finger contact model,and linearizes the contact model,and grabs the point at the fingertip of the known soft hand In this case,optimize the minimum normal contact force of each point;based on the fingertip contact normal force,obtain the torque and bending angle of each finger joint according to the soft hand statics and inverse kinematics,and use the experiment The measured air pressure angle and torque relationship of the soft finger joint airbag model,that is,the load characteristic,obtains the air pressure required to grasp the object.Finally,the above-mentioned theory is applied to the plane grasping of typical objects.The software hand robot grasping simulation technology based on CoppeliaSim is studied.A software hand robot grasping simulation platform was built,and the focus was on target recognition and three-dimensional positioning through vision in a simulation environment to provide guidance for soft hand grasping.The simulation of robotic arm and soft hand grasping was completed.The mechanical arm and soft hand grasping experiment system are designed.Designed a software hand grasping control system experimental plan based on visual target recognition and positioning,introduced the overall design of the robotic arm and software hand grasping system,including the hardware experimental platform and software control platform,and completed the host computer interface and control based on Lab VIEW The system program is compiled;the depth data of the Kinect camera is corrected,the hand-eye calibration of the robotic arm is completed,and the type recognition and three-dimensional positioning of specific targets are realized on the experimental platform,and the specific analysis and comparison of the experimental program are carried out.The accuracy of eye vision positioning and monocular vision positioning,and finally through the control manipulator and soft hand robot to grasp the task of the target,verifies the effectiveness of the theory and plan of the full text.