Vision-based FDM Printing for Fabricating Airtight Soft Actuators

TL;DR

This work addresses the challenge of fabricating airtight soft actuators using desktop FDM by introducing a vision-based closed-loop printing system. The method monitors each printed layer with a camera, projects G-code contours to identify the current layer, detects holes and gaps, and corrects defects in real time through whole-layer ironing of the affected layer. Empirical results show substantial improvements in airtightness across a range of print parameters for TPU filaments, achieving leak-rate reductions up to about 98%. The approach offers a cost-effective pathway to robust, parameter-tolerant fabrication of airtight soft robots, with potential extensions to depth sensing and learning-based defect detection.

Abstract

Pneumatic soft robots are typically fabricated by molding, a manual fabrication process that requires skilled labor. Additive manufacturing has the potential to break this limitation and speed up the fabrication process but struggles with consistently producing high-quality prints. We propose a low-cost approach to improve the print quality of desktop fused deposition modeling by adding a webcam to the printer to monitor the printing process and detect and correct defects such as holes or gaps. We demonstrate that our approach improves the air-tightness of printed pneumatic actuators without fine-tuning printing parameters. Our approach presents a new option for robustly fabricating airtight, soft robotic actuators.

Figures (7)

• Figure 1: Overview of our closed-loop, vision-based printing strategy for FDM. (A) Printing defects that could lead to leakage but cannot be fixed with existing methods. (B) Photos of using the same setup and parameters to print a bellow actuator. When the correction subsystem is not in use, the bellow actuator leaks (left), while the bellow actuator printed with the correction subsystem is fully airtight (right). (C) Concept figure: our closed-loop printing strategy has the potential to improve the airtightness of soft systems in comparison to open-loop printing strategies; we detect and correct defects during the printing process.
• Figure 2: Block diagram of our closed-loop FDM printing system. Green blocks represent hardware interfacing with software subsystems. Print pre-processing: processes the original G-code into a layer-by-layer format, embedding movement commands for image capture, and sends comments to the printer. Detection: identifies defective areas in the current layer through real-time image analysis. Correction: rectifies printing errors within the current layer.
• Figure 3: Defect detection pipeline. For each layer, contours are extracted from the G-Code and mapped onto the captured image. Utilizing these projected boundaries, the most recently printed layer is segmented from the overall image. Subsequently, areas of low intensity are filtered from this segmented portion. Regions falling within a pre-defined size range are classified as defects.
• Figure 4: Z-Shape wiping tower for nozzle cleaning. The Z-shaped structure serves as a transition path for the nozzle, moving from the image capture location to the subsequent printing position. This pathway allows the nozzle to wipe off any oozed filament, ensuring optimal extrusion conditions before resuming printing.
• Figure 5: Microscope image of uncorrected and corrected layers. (A, C) Uncorrected holes within the structure, particularly in thin-walled sections, compromise the airtightness of the structure. (B, D) The correction process employs targeted melting of adjacent materials and supplemental extrusion to effectively seal the holes.