Robot Posture Control for Precise Trimming of Carbon Fiber Preforms

Precise robotic trimming of carbon fiber preforms is essential in resin transfer molding to ensure mold conformity and final product accuracy. However, the low stiffness of industrial robots causes tool center point (TCP) displacement under cutting forces, resulting in dimensional errors. Conventional posture optimization using redundant degrees of freedom (DOF) is ineffective for knife-based trimming in which all six DOFs are constrained. This study proposes a stiffness-based posture control method applicable to nonredundant cutting tasks. Despite the absence of redundancy, the robot can assume eight distinct postures on the same TCP. A virtual joint modeling approach was employed to predict posture-dependent deformation and identify the configuration with maximal stiffness. Experimental validation showed that the selected high-stiffness posture reduced the dimensional error by 64.3% compared with the lowest-stiffness case, demonstrating the potential of the method to enhance the accuracy in fully constrained robotic trimming processes.