Haofeng Chen presents Toward Large-Area Robotic Skin
On 2026-06-11 - 2026-06-11 10:00:00 at G205, Karlovo náměstí 13, Praha 2
Title: Toward Large-Area Robotic Skin: Multitouch, Force, and Bending Sensing
with Electrical Impedance Tomography
Abstract: Robotic skins promise safer and richer physical interaction between
humans and robots, but whole-body tactile sensing remains difficult because
sensors must cover large, curved, and deformable surfaces while preserving
information about contact location, force, and body state. This seminar
presents
two recent approaches based on Electrical Impedance Tomography (EIT) that
address complementary aspects of this challenge. The first work combines an EIT
layer with a pneumatic pressure layer: EIT provides robust multi-contact
localization, while the pneumatic channel supplies a stable force signal that
is
corrected using location-aware calibration fields. This dual-channel strategy
substantially improves force estimation while avoiding heavy EIT-only
calibration pipelines. The second work extends EIT tactile sensing to
deformable
surfaces by introducing adaptive reference management, state classification,
and
regression-based bending angle estimation. Together, the papers show how EIT
can
move beyond static touch localization toward scalable robotic skins capable of
multi-touch perception, force estimation, and deformation-aware sensing for
humanoid and soft robotic systems.
Articles:
Chen, H., Himmel, B., Li, B., Wang, X. and Hoffmann, M. (2026), 'Multi-touch
and
bending perception using electrical impedance tomography for robotics', IEEE
Robotics and Automation Letters. [accepted] https://arxiv.org/abs/2503.13048
Chen, H., Kubik, J., Himmel, B., Hoffmann, M. and Lee, H. (2026), 'Dual-Channel
Tomographic Tactile Skin with Pneumatic Pressure Sensing for Improved Force
Estimation', IEEE Sensors. [accepted] https://arxiv.org/abs/2503.13036
with Electrical Impedance Tomography
Abstract: Robotic skins promise safer and richer physical interaction between
humans and robots, but whole-body tactile sensing remains difficult because
sensors must cover large, curved, and deformable surfaces while preserving
information about contact location, force, and body state. This seminar
presents
two recent approaches based on Electrical Impedance Tomography (EIT) that
address complementary aspects of this challenge. The first work combines an EIT
layer with a pneumatic pressure layer: EIT provides robust multi-contact
localization, while the pneumatic channel supplies a stable force signal that
is
corrected using location-aware calibration fields. This dual-channel strategy
substantially improves force estimation while avoiding heavy EIT-only
calibration pipelines. The second work extends EIT tactile sensing to
deformable
surfaces by introducing adaptive reference management, state classification,
and
regression-based bending angle estimation. Together, the papers show how EIT
can
move beyond static touch localization toward scalable robotic skins capable of
multi-touch perception, force estimation, and deformation-aware sensing for
humanoid and soft robotic systems.
Articles:
Chen, H., Himmel, B., Li, B., Wang, X. and Hoffmann, M. (2026), 'Multi-touch
and
bending perception using electrical impedance tomography for robotics', IEEE
Robotics and Automation Letters. [accepted] https://arxiv.org/abs/2503.13048
Chen, H., Kubik, J., Himmel, B., Hoffmann, M. and Lee, H. (2026), 'Dual-Channel
Tomographic Tactile Skin with Pneumatic Pressure Sensing for Improved Force
Estimation', IEEE Sensors. [accepted] https://arxiv.org/abs/2503.13036