Computational Design of Metal Surfaces and Human-Machine Collaborative Construction: Parameter-Driven Computational Design and Multi-Agent Collaborative Construction Methods for Metal Surfaces

Abstract

The design and formation of freeform metal surfaces are disconnected, manufacturing constraints are difficult to pre-set, and human-machine collaboration efficiency is low. To address these issues, a parametrically excited computational design and multi-agent collaborative formation method for metal surfaces is developed. First, NURBS surfaces are used as the core geometric representation, and process constraints such as curvature, plate thickness, and forming radius are explicitly incorporated into the parameter space, creating a multi-objective improved model that considers geometric approximation, structural performance, and manufacturing costs. Next, a human-machine-measurement multi-agent system architecture is presented. Through task parameterization coding and collaborative scheduling strategies, the improved surface family is transformed into executable assembly tasks. Solid-state verification is conducted using six sets of templates with different curvatures and scales. The results show that when the RMSE is controlled within 4.2 mm, the total project duration is reduced by an average of approximately 22%, robot utilization is significantly improved, and the average RPE of human workers decreases by approximately 2 levels. This demonstrates that the method can ensure geometric accuracy, effectively improve manufacturing efficiency, and reduce the burden on both humans and machines.