A High-Performance Modular Linear Actuator With Enhanced Compensation of Cogging Torque and Friction

This article presents a compact linear actuator featuring a high force-to-weight ratio enabled by integrating all the drive, transmission, and sensing control modules. A four-guide-rod design enhances bending and torsional stiffness with low mass. An iterative method collects and compensates cogging torque of the linear actuator, ensuring reduced speed fluctuations, improved motion stability, and rapid convergence. Position-dependent friction is addressed by a hybrid Stribeck-neural network strategy, which captures residual effects to enhance low-speed smoothness and tracking accuracy. The integrated framework combines model-based and data-driven strengths to deliver precise, robust, and high-performance linear actuation and effectively mitigate dynamic disturbances.