Abstract:In this paper we focus on analyzing the thermal modality of tactile sensing for material recognition using a large materials database. Many factors affect thermal recognition performance, including sensor noise, the initial temperatures of the sensor and the object, the thermal effusivities of the materials, and the duration of contact. To analyze the influence of these factors on thermal recognition, we used a semi-infinite solid based thermal model to simulate heat-transfer data from all the materials in the CES Edupack Level-1 database. We used support-vector machines (SVMs) to predict F1 scores for binary material recognition for 2346 material pairs. We also collected data using a real robot equipped with a thermal sensor and analyzed its material recognition performance on 66 real-world material pairs. Additionally, we analyzed the performance when the models were trained on the simulated data and tested on the real-robot data. Our models predicted the material recognition performance with a 0.980 F1 score for the simulated data, a 0.994 F1 score for real-world data with constant initial sensor temperatures, a 0.966 F1 score for real-world data with varied initial sensor temperatures, and a 0.815 F1 score for sim-to-real transfer. Finally, we present some guidelines on sensor design and parameter choice for thermal recognition based on the insights gained from these results that would hopefully enable robotics researchers to use this less-explored tactile sensing modality more effectively during physical human-robot and robot-object interactions. We release our simulated and real-robot datasets for further use by the robotics community.

Thermal Dissipation Resulting from Everyday Interactions as a Sensing Modality -- The MIDAS Touch

Thermal analysis of ultrathin, compliant sensors for characterization of the human skin

Data-Driven Contact-Based Thermosensation for Enhanced Tactile Recognition

Ultrathin Conformal Devices for Precise and Continuous Thermal Characterization of Human Skin

Multi-sensor Spatial Augmented Reality for Visualizing the Invisible Thermal Information of 3D Objects

Analyzing Material Recognition Performance of Thermal Tactile Sensing using a Large Materials Database and a Real Robot

Hassle-free Approach to Thermal Transport Measurements Using Spatial-Temporal Temperature Data

Thermaterial: Program Ambient Heat Transfer Behaviors on Objects Through Fluidic Composites

Highly Precise Multifunctional Thermal Management-Based Flexible Sensing Sheets

Advanced Approaches for Quantitative Characterization of Thermal Transport Properties in Soft Materials Using Thin, Conformable Resistive Sensors

Heat Flux Sensing for Machine-Learning-Based Personal Thermal Comfort Modeling

Mid-infrared spectroscopic thermotransmittance measurements in dielectric materials for thermal imaging

ThermalRing: Gesture and Tag Inputs Enabled by a Thermal Imaging Smart Ring

A multi-modal approach to continuous material identification through tactile sensing

Thermal Magnifier and Minifier

Rapidly Developing Human Heat Residue Model under Various Conditions Based on Fluent and Thermal Video

Integration of Independent Heat Transfer Mechanisms for Non-Contact Cold Sensation Presentation With Low Residual Heat

Depth Thermography: Noninvasive 3D Temperature Profiling Using Infrared Thermal Emission

Material Recognition via Heat Transfer Given Ambiguous Initial Conditions

Quantitative Thermosensation in a Single Touch

Flexible integrated sensor with asymmetric structure for simultaneously 3D tactile and thermal sensing