Year: 2022
Type: article
Abstract: Tactile sensing plays an important role for robots' perception, but the existing tactile technologies have multiple limitations. Visual-tactile sensor (VTS) is a newly developed tactile detector; it perceives the contacting surface shape, or even more refined texture, by way of the contact deformation image captured by a camera. A conventional visual perception is usually formulated as a data processing. It suffers issues of cumbersome training set and complicated calibration procedures. A novel model-based depth perceptual scheme is proposed where a mapping from the image intensity to the contact geometry is mathematically formulated with an associated tailored fast solver. The hardware calibration requires single image only, leading to an outstanding algorithmic robustness. The non-uniformity of the illumination condition is embodied by the stereo model, resulting in a robust depth perception precision. Compression tests on a prototype VTS showed the method's capability in high-quality geometry reconstruction. Both contacting shape and texture were captured at a root-mean-square error down to a sub-millimeter level. The feasibility of the proposed in a pose estimation application is further experimentally validated. The associated tests yielded estimation errors that were all less than 3° in terms of spatial orientation and all less than 1mm in terms of translation.
Source: Sensors
DOI: 10.3390/s22176470
Authors: Jingjing Ji(Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing), Yuting Liu(Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Writing – original draft, Writing – review & editing), Huan Ma(Conceptualization, Investigation, Writing – review & editing)
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,513
OpenAlex ID: W4293661600
