Research On Key Technologies Of A New Direct-drive Wafer Handling Robot

Wafer handling robots are the key equipment to achieve stable and high-precision positioning and transmission of wafers.At present,most wafer handling equipment in China still adopts the traditional indirect transmission mode of servo motor and belt drive,which not only has many transmission parts,but also has low positioning accuracy.Maintenance costs are high.The direct-drive wafer handling robot can realize lightweight,high-precision,stable and reliable workpiece transfer,and will be the mainstream structure for wafer handling in the future.Due to less research on directdrive wafer handling robots in China,there is still a big gap between their technical level and large-scale industrial applications.At present,the direct-drive robots used in my country's semiconductor industry introduce foreign technologies.So,this paper investigates the current status of the application of direct-drive wafer handling robots,and conducts research on related key technologies,aiming to break the monopoly of foreign technology and improve the technical level of domestic semiconductor manufacturing enterprises.The main research contents are:(1)The transmission scheme and structure type of common industrial robots are analyzed and compared,and the basic configuration of SCARA robot is selected in combination with the wafer handling scene,and the corresponding direct-drive wafer handling robot body is designed.According to the needs of wafer handling,a frameless torque motor and an absolute value magnetoelectric encoder are selected for assembly,which realizes the direct drive and precise position feedback of the wafer handling robot.In the Solid Works environment,the three-dimensional model of the direct-drive wafer handling robot is established,and the electrical framework principle of the robot's multi-axis linkage is given.(2)The kinematics model of the new direct-drive handling robot is analyzed and deduced,and the corresponding Jacobian matrix is established.The quintic polynomial interpolation method is used to realize the trajectory planning of the vertical axis,and the smooth and continuous speed and acceleration are realized;in the direction of plane motion,the particle swarm optimization algorithm is used to perform the 3-5-3polynomial segment interpolation algorithm in the time domain.optimization.Simulation experiments show that under the premise of ensuring smooth transmission,the optimization algorithm can shorten the handling time and improve the efficiency of wafer handling.(3)The dynamic model of the direct-drive wafer handling robot is established.In view of the unpredictable factors such as the flexibility of the joints and the disturbance of model parameters during the wafer handling process,the RBF neural network structure is selected to approximate the uncertain terms in the dynamic model,and the robust terms are introduced to enhance the stability of the control system.The linear weights in the RBF neural network algorithm are studied by simulation method,and the group assignment optimization algorithm is used to improve the trajectory tracking accuracy,response speed and anti-interference ability of the adaptive controller for each joint.(4)For the simulated work scene of wafer handling,the control system architecture of PC + motion controller + servo drive is designed,and the motion control implementation method of the direct-drive wafer handling robot based on ACS controller is studied.On the Windows platform The simulation motion control program of the wafer handling robot is written in C# language,which verifies the validity of algorithms such as kinematics and path planning.