A dataset for tactile textures on uneven surfaces collected using a BioIn-Tacto sensing module

Area of Science:

• Robotics
• Materials Science
• Sensor Technology

Background:

• Human-like robotic manipulation requires effective texture recognition in varied environments.
• Robots with tactile sensors must identify textures using touch-related features in unpredictable settings.
• Advancing texture classification necessitates a comprehensive dataset of physical interactions between robotic probes and textures.

Purpose of the Study:

• To create a novel dataset from signals collected by a bioinspired multimodal tactile sensing module.
• To capture dynamic interactions between a robotic probe and 12 distinct tactile textures.
• To support research in robotic texture classification and manipulation.

Main Methods:

• A bioinspired multimodal tactile sensing module was employed.
• Signals including pressure, acceleration, angular rate, and magnetic field variations were recorded at 130 Hz.
• Each of the 12 textures was explored 25 times using a sliding motion, resulting in 300 total exploratory episodes.

Main Results:

• A comprehensive dataset of tactile texture interactions was generated.
• The dataset includes multi-sensor signals capturing physical contact dynamics.
• The data encompasses 300 exploratory episodes across 12 different textures.

Conclusions:

• The created tactile texture dataset is valuable for object recognition and robotic manipulation.
• This dataset facilitates tactile texture reconstruction and recognition tasks.
• It provides opportunities to study time series properties in robotic tactile exploration.