A point cloud registration algorithm considering multi-allowance constraints for robotic milling of complex parts

Adaptive allocation of the machining allowance is the crucial factor in ensuring the machining accuracy of complex parts. In this work, we present a multi-objective constraint registration method. First, an improved point cloud segmentation method is developed by combining point search and region data expansion algorithms. Afterward, the machining allowance is accurately calculated by using statistical analysis and multi-point sampling strategies to enhance the calculation accuracy of the point-to-triangular patch distance. Finally, a registration objective function is established by considering the allowance constraints of various geometric regions of the workpiece, and the particle swarm optimization algorithm is used to solve the optimum solution. The proposed multi-constraint registration method realizes optimal allocation of the allowance in different regions, which offers a reference coordinate system for the robotic milling of complex free-formed parts. Simulation and experimental results reveal that the developed method satisfies the minimum registration error while ensuring the allocation of allowance in the robotic milling of the casing cavity compared with other methods.