Research On The Design And Control Of Anthropomorphic Dexterous Hand

With the rapid development of robot technology,the fields of robot application have been expanding.Various working environments and complex work tasks have placed higher demands on robots.The anthropomorphic dexterous hand is an important end-executing mechanism of the intelligent robot,which directly determines the task execution level of the robot.Its versatility and flexibility make it a research hotspot in the field of robotics.An anthropomorphic five-finger dexterous hand is designed with reference to the structure and movement characteristics of the human hand.The dexterous hand has 12 active degrees of freedom and is driven by 12 built-in motors.The fingers are modular-designed and the four-link mechanism is used to couple the movement of the DIP and PIP joints,and the position of the finger is optimized with bionics.The dexterous hand was fabricated by SLA 3D printing technology,and its electromechanical system is also completed,including the design of the drive transmission system,sensors,and control circuits,etc.In order to conveniently control and experiment,the host computer control software was also developed based on Qt.The forward kinematics and inverse kinematics of the dexterous hand were analyzed,and the trajectory and motion space of the finger were analyzed by Robotic Toolbox.Statics and dynamic equations of the dexterous hand is derived.Then the analysis of the kinematics of the entire dexterous hand establishes the theoretical basis for the control of the dexterous hand.The position control of the finger is realized by the combination of second-order nonlinear tracking-differentiator(TD)and PID control.The simulation is carried out with Simulink and the joint position tracking experiment and trajectory tracking experiment are carried out to verify the accuracy of the control algorithm.Finally,based on the image recognition technology,the Otsu is used to extract the human hand gesture and split the background according to the skin color.The position and number of fingers,palm position,angle between fingers and other key geometric features are dynamically extracted to control the robot to realize human-computer interaction.