Adaptable Method for Dynamic Planning of 3D Spatial Wireframes Extrusion with Neural Networks and Robotic Automation.

3D extrusion of spatial wireframe provides an alternate method for additive manufacture, which holds a significant advantage over the typical layering printing method in terms of material and time efficiency. However, the complexity of the structure and unsupported printing process are major challenges for path planning of extrusion process. This article presents a method to plan and control the path for extruding 3D spatial wireframes by synthesizing dynamic material behaviors such as heat deformation, plasticity, and bending, during the curing process as active contributors to the printed form. With this method, a novel system of printing spatial wireframe model is developed that disassociate the extruding path and extruded form by dimensional, that is, printing a 3D spatial wireframe with a 2D printing path. The workflow of the research combines robotic/mechanical automation with machine vision and artificial intelligence, to generate material models without any preexisting knowledge of the material. Based on the feedback loop between machine vision and printing control, this system is capable of automatically conducting material experiments at a large scale, observe the results, and learn to generate an end-to-end solution that directly bridges the design intention to the fabrication of spatial wireframes based on nonstandard material behaviors. The development of the method allows designers to design and fabricate dynamic spatial frames, without the support of existing material and structural models.