Near‐Sensor Reservoir Computing for Gait Recognition via a Multi‐Gate Electrolyte‐Gated Transistor

Abstract

The recent emergence of various smart wearable electronics has furnished the rapid development of human-computer interaction, medical health monitoring technologies, etc. Unfortunately, processing redundant motion and physiological data acquired by multiple wearable sensors using conventional off-site digital computers typically result in serious latency and energy consumption problems. In this work, a multi-gate electrolyte-gated transistor (EGT)-based reservoir device for efficient multi-channel near-sensor computing is reported. The EGT, exhibiting rich short-term dynamics under voltage modulation, can implement nonlinear parallel integration of the time-series signals thus extracting the temporal features such as the synchronization state and collective frequency in the inputs. The flexible EGT integrated with pressure sensors can perform on-site gait information analysis, enabling the identification of motion behaviors and Parkinson's disease. This near-sensor reservoir computing system offers a new route for rapid analysis of the motion and physiological signals with significantly improved efficiency and will lead to robust smart flexible wearable electronics.

Reconfigurable optoelectronic transistors for multimodal recognition

“…The rise and decay of the channel currents reflect the temporal characteristics of the optical pulse stimulation, which is indicative of the device’s high suitability for RC (Supplementary Fig. 7 ) 34 . Furthermore, the response under blue light (450 nm) illumination was investigated, and a similar volatile behavior as that under low-intensity UV illumination was observed, but with a much smaller magnitude of change (Supplementary Fig.…”

“…Electrolyte-gated transistors (EGT), with tunable ion dynamic timescales at different gate voltages, show great potential for processing dynamic information in a single device 33 , 34 . Our previous study demonstrated that the reconfigurable property derives from the electric double layer (EDL) and the ion migration mechanisms 35 .…”

Harness Background Lighting for Separability Enhancement in Optical Sensing Reservoir for Temporal Signal Processing Utilizing Amorphous IGZO Neuromorphic Transistors

“…As a result, reservoir computing (RC) has emerged as a promising algorithm for processing temporal data. − In the RC, temporal signals are transformed into high-dimensional states, enabling complex inputs to become linearly separable on the basis of reservoir states. This transformation of inputs enables efficient processing through a simple readout layer, resulting in significant reductions in network scale and training cost. , Recently, several physical reservoirs have been proposed, including those utilizing metal oxides, ,, two-dimensional materials, − and organics. − The effectiveness of the RC system, which is characterized by fading memory and distinct reservoir states, has been successfully demonstrated in these studies. Nonetheless, only a few of these reservoirs possess the combined features of light responsiveness, CMOS compatibility, and scalability.…”

Research on tactile sensation by physical reservoir computing with a robot arm and a Ag2S reservoir

“…Various physical reservoirs have been used in the reservoir layer, [11][12][13][14][15][16][17] and research into materials-based reservoir computing is currently accelerating. [18][19][20] For example, reservoir computing research exists using atomic switch technologies to control the growth and contraction of Ag filaments from Ag 2 S. 12,21,22) Ag 2 S reservoirs have also been used for benchmarking tasks such as STM, PC, and MNIST image recognition and optical pattern classification. 23,24) In terms of social implementation of artificial intelligence, AI robot development using machine learning, such as deep learning, has been conducted in recent years, [25][26][27] which will meet the increasing demand for care and assistance robots.…”