Brain inspired method boosts tactile sensing across unknown sensors
BIFTA: Brain-Inspired Few-Shot Tactile Adaptation for Unknown Sensors
RoboticsArtificial Intelligence
Summary
Tactile sensors help robots feel and understand objects, but models trained on one type of sensor often fail when used with new, unknown sensors. The authors propose BIFTA, a method inspired by how the brain quickly adapts to new sensory inputs. BIFTA adjusts a fixed tactile sensor model using only a few labeled examples from the new sensor, improving accuracy drastically. Their experiments show that BIFTA works well across different datasets and tasks, enabling better tactile sensing on diverse hardware.
What this means in practice
- •For robotics developers: Improve tactile perception systems by quickly adapting models to new tactile sensors with minimal labeled data.
- •For industrial automation teams: Enable flexible deployment of robotic grippers with different tactile hardware without retraining models from scratch.
Authors
Boheng Liu, Ziyu Li, Xia Wu
Abstract
Advances in tactile sensing have made contact-rich perception possible, accelerating progress in robotic manipulation, material understanding, and embodied interaction. However, because optical design, elastomer mechanics, and imaging geometry differ substantially across tactile sensors, models trained on known sensor types can suffer an abrupt performance collapse on unknown sensors. To address this problem, we propose the Brain-Inspired Few-Shot Tactile Adaptation (BIFTA) framework; it draws on the brain's rapid sensory adaptation mechanism to adapt a frozen encoder to an unknown tactile sensor from a small labeled support set. BIFTA preserves pretrained representations through dual-view statistical memory, constructs support-conditioned spectral graphs to repair sensor-dependent feature neighborhoods, and applies uncertainty-gated recurrent propagation to strengthen reliable cross-query evidence. Extensive benchmarks across three tactile datasets show that BIFTA substantially improves adaptation to unknown sensors: with only 10\% labeled target data on SITR, it raises mean Sparsh accuracy from 6.86\% for the frozen source classifier to 87.09\%, exceeding the strongest implemented prior comparison by 47.22 percentage points, and these gains generalize across datasets, pretrained backbones, and tactile tasks. These results validate BIFTA for data-efficient adaptation to unknown tactile sensors and offer a promising route toward tactile models that transfer across heterogeneous hardware.