An Artificial Sensory Neuron with Tactile Perceptual Learning

Abstract

Sensory neurons within skin form an interface between the external physical reality and the inner tactile perception. This interface enables sensory information to be organized identified, and interpreted through perceptual learning-the process whereby the sensing abilities improve through experience. Here, an artificial sensory neuron that can integrate and differentiate the spatiotemporal features of touched patterns for recognition is shown. The system comprises sensing, transmitting, and processing components that are parallel to those found in a sensory neuron. A resistive pressure sensor converts pressure stimuli into electric signals, which are transmitted to a synaptic transistor through interfacial ionic/electronic coupling via a soft ionic conductor. Furthermore, the recognition error rate can be dramatically decreased from 44% to 0.4% by integrating with the machine learning method. This work represents a step toward the design and use of neuromorphic electronic skin with artificial intelligence for robotics and prosthetics.

Bioinspired Prosthetic Interfaces

“…The synaptic transistors integrated with pressure sensors offer the possibility to recognize and differentiate features of patterns through machine learning by sliding the sensor over the surface of the object . Wan et al developed a neuromorphic tactile processing system (NeuTap), which is comprised of a piezoresistive sensor, a soft ionic conductor, and an ion‐gated transistor.…”

High‐Performance Flexible Bioelectrocatalysis Bioassay System Based on a Triphase Interface

“…Flexible devices have attracted extensive research interest in the fields of physiological monitoring, intelligent display, energy collection, and storage . Flexible biosensors based on individual's physiological biomarkers have recently been fabricated that are able to provide insight into the user's health state at a molecular level as well as diabetic monitoring .…”

Parallel Signal Processing of a Wireless Pressure‐Sensing Platform Combined with Machine‐Learning‐Based Cognition, Inspired by the Human Somatosensory System

“…However, tactile sensing technology alone is insufficient to enable the aforementioned applications, where the processing of acquired signal is also of critical importance . In fact, only when tactile sensing technology is combined with the correspondingly fitting signal processing technology, will various devices be able to perceive and interact with the environment, and hence be able to “feel like a human.” In this regard, one could learn and gain inspiration from the human somatosensory system, which is a system of receptors, sensory neurons, and synaptic pathways, through which our body receives and processes tactile information . Our somatosensory system has the following interesting features to consider in terms of tactile signal processing ( Figure a).…”

“…In recent years, researchers have begun to mimic the human somatosensory system, where tactile sensors were combined with artificial synaptic devices, by which multiple tactile signals were combined and processed simultaneously . For instance, Lee et al made a multi‐device system that can convert incoming pressure signals into electrical pulses using ring oscillators, which was then processed by synaptic devices .…”

“…However, it is not possible to ascertain which pressure sensor a given signal is coming from; hence, additional synaptic devices are needed. In another report, two pressure sensors were connected to a single synaptic device where the signals from both pressure sensors were processed simultaneously, and through learning, recognition error was progressively reduced . However, as pressure sensors connected to a synaptic device increase, it will become increasingly complex to differentiate the pressure levels of each sensor.…”