Lumped-Parameter Modeling and Control for Robotic High-Viscosity Fluid Deposition

William van den Bogert,James Lorenz,Xili Yi,Albert J. Shih,Nima Fazeli
DOI: https://doi.org/10.1109/lra.2024.3349931
IF: 5.2
2024-02-01
IEEE Robotics and Automation Letters
Abstract:Robotic high-viscosity fluid deposition plays a pivotal role in various manufacturing applications including adhesive and sealant dispensing, as well as in the additive manufacturing of deformable materials, such as those employed in soft robotics. Uncompensated high-viscosity fluid deposition can lead to poor part quality and defects due to large transient delays and complex fluid dynamics. In this letter, we propose a lumped-parameter flow model and compensation strategies to address significant transient delays and nonlinearity inherent in high-viscosity fluid deposition using a robotic manipulator. Our computationally efficient model is well-suited to real-time control and can be calibrated in minutes. Our compensation strategies leverage an iterative Linear-Quadratic Regulator to compute compensated deposition paths that can be deployed on robotic dispensing systems. These paths can either be deployed offline or corrected live via feedback from our proposed vision-based flow sensor. To validate the effectiveness of our approach, we conducted experiments extruding high-viscosity liquid silicone using a Kuka lbr iiwa robot. Comparative analysis with several baseline protocols demonstrates that our proposed method significantly improves material deposition within desired boundaries.
robotics
What problem does this paper attempt to address?