Abstract
Soft, sensorized companions provide a physically safe and emotionally intuitive interface, yet their deformability and multichannel tactile sensing complicate robust interpretation of human affect. This study presents a complete open-source MATLAB-based framework for developing and validating compact deep learning models for affective touch recognition in soft interactive companions.
A primary contribution is the public availability of a diverse FAIR-compliant dataset of 1326 labeled gesture sequences collected from 25 participants, spanning children, teenagers, and adults, providing a reusable resource for future research in affective touch recognition. Systematic architecture and hyperparameter exploration across 468 CNN models identified compact dilated one-dimensional convolutional neural networks (1D CNNs) as the most effective solution, with a 13.2k-parameter model achieving 75% test accuracy and 85% mean leave-one-subject-out cross-validation accuracy.
Theoretical inference-time analysis shows that quantized deployment requires 3.2 MMAC per window, compatible with 20 Hz real-time operation on the target microcontroller. PC-based real-time simulation with physical toy streaming sensor data demonstrates that the CNN resolves subtle social touches that the previous heuristic system failed to detect, whereas high-force negative interactions are captured more reliably by trivial threshold-based logic. The resulting hybrid inference pipeline - instantaneous heuristic filtering followed by CNN-based nuanced gesture classification - is proposed as the embedded deployment strategy.
This study demonstrates that emotionally meaningful, privacy-preserving touch interpretation is computationally feasible for direct embedding within soft therapeutic companions, with hardware integration addressed in a forthcoming study.
Blogger's Review: This paper significantly advances the field of affective touch recognition by providing an open-source framework. The performance of the lightweight 1D CNN in real-time applications is impressive, especially in resolving subtle social touches, showcasing the potential of deep learning in social assistive technologies. Future research could further explore hardware integration possibilities for broader applications.