Soft-bodied robots with multimodal sensing capabilities hold promise for versatile and user-friendly robotics. However, seamlessly integrating multiple sensing functionalities into soft artificial skins remains a challenge due to compatibility issues between soft materials and conventional electronics. While vision-based tactile sensing has enabled simple and effective sensor designs for robotic touch, there has been limited exploration of this technique for intrinsic multimodal sensing in large-sized soft robot bodies. To address this gap, this paper introduces a novel vision-based soft sensing technique, named ProTac, capable of operating either in tactile or proximity sensing modes. This vision-based sensing technology relies on a soft functional skin that can actively switch its optical properties between opaque and transparent states. Furthermore, the paper develops efficient learning pipelines for proximity and tactile perceptions, as well as sensing strategies enabled through the timing activation of the two sensing modes. The effectiveness of the soft sensing technology is demonstrated through a soft ProTac link, which can be integrated into newly constructed or existing commercial robot arms. Results suggest that robots integrated with the ProTac link, along with rigorous control formulation can perform safe and purposeful control actions, which enhances human-robot interaction scenarios and facilitates motion control tasks that are challenging to achieve with conventional rigid links. Supplementary video:https://youtu.be/dFgZLUpeWw4Project website:https://quan-luu.github.io/protac-website/

Large-scale robotic skin with tactile sensing ability is emerging with the potential for use in close-contact human–robot systems. Although recent developments in vision-based tactile sensing and related learning methods are promising, they have been mostly designed for small-scale use, such as by fingers and hands, in manipulation tasks. Moreover, learning perception for such tactile devices demands a huge tactile dataset, which complicates the data collection process. To address this, this study introduces a multiphysics simulation pipeline, calledSimTacLS, which considers not only the mechanical properties of external physical contact but also the realistic rendering of tactile images in a simulation environment. The system utilizes the obtained simulation dataset, including virtual images and skin deformation, to train a tactile deep neural network to extract high-level tactile information. Moreover, we adopt a generative network to minimize sim2real inaccuracy, preserving the simulation-based tactile sensing performance. Last but not least, we showcase this sim2real sensing method for our large-scale tactile sensor (TacLink) by demonstrating its use in two trial cases, namely, whole-arm nonprehensile manipulation and intuitive motion guidance, using a custom-built tactile robot arm integrated with TacLink. This article opens new possibilities in the learning of transferable tactile-driven robotics tasks from virtual worlds to actual scenarios without compromising accuracy.