An Adaptive Robot Trajectory Planning Method for Measurement of Thin-Walled Workpieces with Variable Curvature

To address the intelligent detection requirements during the roll-bending process of large aerospace thin-walled workpieces, this paper proposes a robot trajectory planning method for measurement that incorporates dynamic curvature characteristics, aiming to enhance the precision of laser-based inspection. Firstly, an intelligent measurement system is constructed to analyze the kinematic relationships among the thin-walled workpiece, the industrial robot, and the laser camera. A unified coordinate system is established through spatial coordinate transformation. Next, an adaptive sampling strategy is designed based on the curvature distribution of the workpiece, where dynamic curvature thresholds segment the surface cross-sectional profiles. Sampling points are dynamically generated within each subregion according to the laser camera's field of view. Subsequently, Principal Component Analysis (PCA) is employed to calculate surface normal vectors, and these sampling points are transformed into robot trajectory points. To ensure motion smoothness and stability, S-curve algorithm is implemented for joint trajectory planning. Experimental results demonstrate that the proposed method adaptively generates robot trajectories by integrating surface characteristics of aerospace thin-walled workpieces, achieving improvement in measurement accuracy compared to other sampling methods while maintaining robot motion stability.