Human-machine Interface Applications Research Articles

Development and validation of a machine learning‐based model of ischemic stroke risk in the Chinese elderly hypertensive population

AbstractMachine learning (ML) has made some significant contributions to stroke prevention, but the stability and accuracy of existing models for clinical applications are uncertain. This study develops and validates an interpretable ML model using metabolic and coagulation biomarkers to predict ischemic stroke in elderly hypertensive patients in Northwest China. The prediction model used 453 electronic medical records for the model building (80% as a training set and 20% as a test set) and 132 for external validation. The final seven key features (D‐dimer, cystatin C, homocysteine, hemoglobin A1c, prothrombin time, low‐density lipoprotein C, and triglyceride glucose‐body mass index) were selected by the advanced approach, elastic net, and classical wrapping approaches. The final model, eXtreme gradient boosting, was identified as having superior performance than the other 9 classifers (random forest, Gaussian process, multilayer perceptron, logistic regression, support vector machine, K‐nearest neighbor, decision tree, Gaussian naive bayes, and ensemble model), with area under the receiver‐operating characteristic curves of 0.97 and 0.94 for the test and external validation sets, respectively. The final model demonstrates excellent stability, accuracy, and clinical usefulness through various metrics and decision curve analysis. Additionally, an online human–machine interface application has been developed for clinical practice to help early identification and intervention for ischemic stroke in elderly hypertensive patients.

Ultrahighly Sensitive Flexible Pressure Sensors with Dual-Mode Response Based on Monolayer Films of Calcium Niobate Nanosheets.

Flexible pressure sensors present enormous potential for applications in health monitoring, human-machine interfacing, and electronic skins (e-skin). However, at the cost of flexibility, the design of flexible pressure sensors has been facing the trade off between high sensitivity and wide sensing range. Herein, we propose a sandwiched structure composed of monolayer films of calcium niobate nanosheets to endow the device with both ultrahigh sensitivity and a wide sensing range. Tunable contact between the two electrodes of the pressure sensor through the gaps in the insulative monolayer film and precise thickness modulation of the monolayer films at the nanoscale result in an ultrahigh sensitivity and wide sensing range of the sensors. By virtue of these two traits, the pressure sensor distinguishes itself with unprecedented performances of ultrahigh sensitivity (6.43 × 104 kPa-1), a wide sensing range (1.94-60.00 kPa), a fast response time (<165 ms), and reliable repeatability. In addition, the sensor has a sensing mechanism transition from capacitive mode to piezoresistive mode from low pressure to high pressure. The sensors demonstrates the ability for motion detection of the human body. This work sheds light on the development of highly sensitive flexible pressure sensors.

Air-Writing Recognition Enabled by a Flexible Dual-Network Hydrogel-Based Sensor and Machine Learning.

Accurate air-writing recognition is pivotal for advancing state-of-the-art text recognizers, encryption tools, and biometric technologies. However, most existing air-writing recognition systems rely on image-based sensors to track hand and finger motion trajectories. Additionally, users' writing is often guided by delimiters and imaginary axes which restrict natural writing movements. Consequently, recognition accuracy falls short of optimal levels, hindering performance and usability for practical applications. Herein, we have developed an approach utilizing a one-dimensional convolutional neural network (1D-CNN) algorithm coupled with an ionic conductive flexible strain sensor based on a sodium chloride/sodium alginate/polyacrylamide (NaCl/SA/PAM) dual-network hydrogel for intelligent and accurate air-writing recognition. Taking advantage of the excellent characteristics of the hydrogel sensor, such as high stretchability, good tensile strength, high conductivity, strong adhesion, and high strain sensitivity, alongside the enhanced analytical ability of the 1D-CNN machine learning (ML) algorithm, we achieved a recognition accuracy of ∼96.3% for in-air handwritten characters of the English alphabets. Furthermore, comparative analysis against state-of-the-art methods, such as the widely used residual neural network (ResNet) algorithm, demonstrates the competitive performance of our integrated air-writing recognition system. The developed air-writing recognition system shows significant potential in advancing innovative systems for air-writing recognition and paving the way for exciting developments in human-machine interface (HMI) applications.

Customized surface adhesive and wettability properties of conformal electronic devices.

Conformal and body-adaptive electronics have revolutionized the way we interact with technology, ushering in a new era of wearable devices that can seamlessly integrate with our daily lives. However, the inherent mismatch between artificially synthesized materials and biological tissues (caused by irregular skin fold, skin hair, sweat, and skin grease) needs to be addressed, which can be realized using body-adaptive electronics by rational design of their surface adhesive and wettability properties. Over the past few decades, various approaches have been developed to enhance the conformability and adaptability of bioelectronics by (i) increasing flexibility and reducing device thickness, (ii) improving the adhesion and wettability between bioelectronics and biological interfaces, and (iii) refining the integration process with biological systems. Successful development of a conformal and body-adaptive electronic device requires comprehensive consideration of all three aspects. This review starts with the design strategies of conformal electronics with different surface adhesive and wettability properties. A series of conformal and body-adaptive electronics used in the human body under both dry and wet conditions are systematically discussed. Finally, the current challenges and critical perspectives are summarized, focusing on promising directions such as telemedicine, mobile health, point-of-care diagnostics, and human-machine interface applications.

A Cost-Effective and Easy-to-Fabricate Conductive Velcro Dry Electrode for Durable and High-Performance Biopotential Acquisition.

Compared with the traditional gel electrode, the dry electrode is being taken more seriously in bioelectrical recording because of its easy preparation, long-lasting ability, and reusability. However, the commonly used dry AgCl electrodes and silver cloth electrodes are generally hard to record through hair due to their flat contact surface. Claw electrodes can contact skin through hair on the head and body, but the internal claw structure is relatively hard and causes discomfort after being worn for a few hours. Here, we report a conductive Velcro electrode (CVE) with an elastic hook hair structure, which can collect biopotential through body hair. The elastic hooks greatly reduce discomfort after long-time wearing and can even be worn all day. The CVE electrode is fabricated by one-step immersion in conductive silver paste based on the cost-effective commercial Velcro, forming a uniform and durable conductive coating on a cluster of hook microstructures. The electrode shows excellent properties, including low impedance (15.88 kΩ @ 10 Hz), high signal-to-noise ratio (16.0 dB), strong water resistance, and mechanical resistance. After washing in laundry detergent, the impedance of CVE is still 16% lower than the commercial AgCl electrodes. To verify the mechanical strength and recovery capability, we conducted cyclic compression experiments. The results show that the displacement change of the electrode hook hair after 50 compression cycles was still less than 1%. This electrode provides a universal acquisition scheme, including effective acquisition of different parts of the body with or without hair. Finally, the gesture recognition from electromyography (EMG) by the CVE electrode was applied with accuracy above 90%. The CVE proposed in this study has great potential and promise in various human-machine interface (HMI) applications that employ surface biopotential signals on the body or head with hair.

Quasi-1D Conductive Network Composites for Ultra-Sensitive Strain Sensing.

Highly performance flexible strain sensor is a crucial component for wearable devices, human-machine interfaces, and e-skins. However, the sensitivity of the strain sensor is highly limited by the strain range for large destruction of the conductive network. Here the quasi-1D conductive network (QCN) is proposed for the design of an ultra-sensitive strain sensor. The orientation of the conductive particles can effectively reduce the number of redundant percolative pathways in the conductive composites. The maximum sensitivity will reach the upper limit when the whole composite remains only "one" percolation pathway. Besides, the QCN structure can also confine the tunnel electron spread through the rigid inclusions which significantly enlarges the strain-resistance effect along the tensile direction. The strain sensor exhibits state-of-art performance including large gauge factor (862227), fast response time (24ms), good durability (cycled 1000 times), and multi-mechanical sensing ability (compression, bending, shearing, air flow vibration, etc.). Finally, the QCN sensor can be exploited to realize the human-machine interface (HMI) application of acoustic signal recognition (instrument calibration) and spectrum restoration (voice parsing).

Flexible Pressure Sensors Based on P(VDF-TrFE) Films Incorporated with Ag@PDA@PZT Particles.

Films of piezoelectric poly(vinylidene fluoride) (PVDF) and its copolymer P(VDF-TrFE) have been studied intensively for their potential application in piezoelectric sensing devices. The present work focuses on tuning the piezoelectric properties of P(VDF-TrFE) films via incorporating Ag and polydopamine co-decorated PZT (Ag@PDA@PZT) particles. Ag@PDA@PZT particles can effectively improve the β-phase content and piezoelectric properties of P(VDF-TrFE) films. The highest open-circuit voltage and short-circuit current of P(VDF-TrFE)-based flexible pressure sensors incorporating Ag@PDA@PZT particles are ~30 V and ~2.4 μA, respectively. The flexible sensors exhibit a response to different body movements, providing a practical and potentially useful basis for human-machine interface applications.

Self-powered and degradable humidity sensors based on silk nanofibers and its wearable and human–machine interaction applications

Humidity sensors have received increasing attention due to their broad potential application in wearable devices, health monitoring, intelligent control, and other fields. However, traditional humidity sensors are prone to produce various toxic substances to pollute the environment after discarding. Also, they are prone to damage during use, which shortens their service life. Therefore, the development of eco-friendly, degradable, and repairable humidity sensors is of great significance for reducing e-waste and protecting the environment. Here, we started with the selection of materials by using silk nanofiber (SNF) film as the substrate material, copper/aluminum as electrodes, and calcium chloride as the electrolyte, combined with the working mechanism of the primary battery, to construct a self-powered humidity sensor. The humidity sensor exhibits excellent humidity sensing performance and good linearity within the range of 30–90 % relative humidity (RH). A single humidity sensor can output 346 mV at 90 % RH. After being damaged, the humidity sensor can be repaired with the help of water. When the ruptured SNF film encounters water molecules, the water molecules can prompt the hydrogen bonds between the SNFs to rearrange and connect, connecting the broken area. Young’s modulus retention rate of the repaired SNF film is over 85 %. The repaired humidity sensor can also output 328 mV. Additionally, an alkaline solution can completely degrade the humidity sensor in less than 80 min. The proposed humidity sensors have been applied to wearable self-powered wristbands, wireless monitoring systems, and non-contact human–machine interaction systems. Hence, we provide comprehensive research from the material level (SNFs) to the functional device level (the self-powered humidity sensor), and then to the system level (wearable and human–machine interface system). The research will open a new way for the design of the self-powered humidity sensor and have great potential applications in wearable electronics and human–machine interface applications.

Photo-programmable hydrogel iontronics for electrically and chromatically rewritable circuits

Hydrogel-based iontronics is emerging as a promising frontier in healthcare and human-machine interfacing (HMI), offering excellent compatibility with biological systems in terms of electrical, chemical, and mechanical properties. However, conventional hydrogel systems have limitations in dynamically regulating their electrical and optical properties, which restricts their use in adaptive electronics and responsive interfaces. In this study, we present a new hydrogel system with UV photochemistry-induced reversible conductivity, enabling reversible changes in conductivity. Unlike typical photo-responsive hydrogels that revert to their original states upon removal of the light source, the new hydrogel can maintain its activated states without continuous light exposure, facilitating practical applications. By leveraging the photobase triphenylmethane leucohydroxide and photoacid n-nitrobenzaldehyde, we achieve a significant increase in photo-induced conductivity compared to existing photo-ionic hydrogels. Combining the effective photo-induced conductivity and the accompanied photochromatic effect, we demonstrate a full hydrogel-based stylus pad capable of tracking motion and strokes, and a soft calculator keypad with programmable conductivity and imprinted patterns. These advancements underscore the importance of actively controlling localized conductivity and processing light inputs in hydrogels, exhibiting their potential for diverse applications in bioelectronics and HMI.

BaZnF4/Polydimethylsiloxane Composite Film-Based Nanogenerators for Energy Harvesting and Human–Machine Interface Applications

BaZnF₄/Polydimethylsiloxane Composite Film-Based Nanogenerators for Energy Harvesting and Human–Machine Interface Applications

Recent progress of optical tactile sensors: A review

Over the past decade, flexible electronics has become a cutting-edge technology owing to its broad applications in robotics, health monitoring, and human-machine interface. The recent development of flexible electrical tactile sensor triggers the research of optical tactile sensors due to their potentials in high-sensitivity, high-precision, low power consumption, and anti-electromagnetic interference. Here, this review mainly discusses the advances and applications of silica fiber, polymer fiber, hydrogel, and micro/nanofiber based tactile sensors, specifically, focusing on the latest research results and on the necessary technologies to enable the health monitoring, robotics, and human-machine interface applications. Furthermore, future outlooks in developing skin-like optical tactile sensors are also presented.

A Triboelectric Nanogenerator-Based Wide Range Self-Powered Flexible Pressure Sensor

Self-powered flexible pressure sensors are needed in applications such as human-machine interaction. Here, we present a triboelectric nanogenerator (TENG) based self-powered flexible pressure sensor capable of operating over a wide-range of pressures (3.2 - 1176 kPa). The sensor exhibits excellent sensitivities in various regions of the pressure range, i.e., 3.16, 0.023 and 0.031 V/kPa in the low (1-10 kPa), medium-high (10-500 kPa) and ultra-high (>500 kPa) pressure regimes respectively. Stable and repeatable TENG responses are achieved for all three pressure regimes, and this is due to the way the real contact area at the TENG interface varies with contact pressure. These results highlight the potential for TENGs in a wide range of applications such as: detecting pressure in wearable devices, human-machine interface, biomedical and automotive applications. As a proof-of-concept, we have demonstrated the use of the presented device in applications such as detection of human and robot finger tapping, collection of human gait information, and detection of impact forces.

Highly stretchable and strain sensitive MXene/MXene:MWCNTs@TPU fiber with hierarchical conductive layers and porous elastic core structure

Fiber flexible strain sensors have the characteristics of small volume, light weight, good flexibility, easy deformation/recovery, and easy weaving. They can conform to various complex surfaces and endure various deformations, such as twisting and bending. Furthermore, the simple preparation process of fiber sensors lowers the production cost significantly. These features make them promising candidates for applications in motion monitoring, human-machine interface, and flexible intelligent products. However, the low sensitivity and narrow sensing range of existing sensors are still major challenges for their further development. To address these challenges, this paper proposes a MXene/MXene:MWCNTs@TPU fiber sensor, which adopts a two-step pre-stretching strategy to modify the porous TPU fiber surface with MXene and MWCNTs. The graded porous TPU fibers obtained by wet spinning can achieve 805 % stretching, providing a reliable substrate support for the fiber sensor, the combination of multi-dimensional conductive materials and the introduction of cracks and wrinkles enhance the sensitivity 277 and sensing range 510 % of the fiber sensor. The fiber sensor can detect various human motions and gestures by directly attaching to the skin or weaving into textiles, and can also be applied to information encryption transmission, demonstrating its huge potential in human health monitoring, intelligent flexible electronic products and encrypted communication.

A Low‐Cost and Do‐It‐Yourself Pressure Sensor Enable Human Motion Detection and Human–Machine Interface Applications

AbstractIn this work, cost‐effective and do‐it‐yourself capacitive pressure sensors are fabricated using readily available commercial components. The sensors are created in a single‐step process ‐by simply applying electrically conductive paint onto both sides of a porous melamine sponge. These sensors exhibit a wide‐range pressure sensing capability, spanning from 10 Pa to 100 kPa. The sensors showcase an impressively low limit of detection, detecting pressures as low as 10 Pa, and exhibit a moderate response time of 123 ms. Moreover, the sensors display remarkable repeatability and stability over 10 000 loading and unloading cycles without experiencing fatigue. Notably, these exceptional qualities come at an exceptionally low material cost, with the sensor measuring 20 × 20 × 2 mm. To showcase their potential applications, the fabricated sensors are successfully employed in real‐time human motion detection, proximity detection, and wearable keyboard applications.

Novel near E-Field Topography Sensor for Human-Machine Interfacing in Robotic Applications.

This work investigates a new sensing technology for use in robotic human-machine interface (HMI) applications. The proposed method uses near E-field sensing to measure small changes in the limb surface topography due to muscle actuation over time. The sensors introduced in this work provide a non-contact, low-computational-cost, and low-noise method for sensing muscle activity. By evaluating the key sensor characteristics, such as accuracy, hysteresis, and resolution, the performance of this sensor is validated. Then, to understand the potential performance in intention detection, the unmodified digital output of the sensor is analysed against movements of the hand and fingers. This is done to demonstrate the worst-case scenario and to show that the sensor provides highly targeted and relevant data on muscle activation before any further processing. Finally, a convolutional neural network is used to perform joint angle prediction over nine degrees of freedom, achieving high-level regression performance with an RMSE value of less than six degrees for thumb and wrist movements and 11 degrees for finger movements. This work demonstrates the promising performance of this novel approach to sensing for use in human-machine interfaces.

High-precision strategy for piezoelectric characterization of nano/microwire

Investigating piezoelectricity of nano/microwire is important for their applications in nanogenerator, actuator, sensor and many more. Up to now, accurately probing the piezoelectric constant of nano/microwire along its axial direction has remained a challenge. Here, we propose a strategy that can high-precisely characterize the piezoelectric constant of nano/microwire along its axial direction through constructing a lateral structure with short connection of two ends of nano/microwire as one electrode and probe tip as another electrode, making the nano/microwire not buckle and its length unchanged during measurements. Theoretical analysis and simulations show that this strategy has ultra-high precision, in which the error can be as low as 2% (aspect ratio ∼100), 0.3% (∼500) and less than 0.2% (∼1000). Using this strategy, the measured piezoelectric constant d33 of ZnO microwires is 12.31 ± 0.57 pm/V, highly consistent with widely accepted values in ZnO bulks and films, further verifying the correctness and accuracy of this strategy. This simple, universal and high-precision strategy contributes to the piezoelectric characterization of nano/microwires and their applications in energy harvesting, human-machine interfacing, active electronics/optoelectronics and many more.

Multi-Channel Soft Dry Electrodes for Electrocardiography Acquisition in the Ear Region.

In-ear acquisition of physiological signals, such as electromyography (EMG), electrooculography (EOG), electroencephalography (EEG), and electrocardiography (ECG), is a promising approach to mobile health (mHealth) due to its non-invasive and user-friendly nature. By providing a convenient and comfortable means of physiological signal monitoring, in-ear signal acquisition could potentially increase patient compliance and engagement with mHealth applications. The development of reliable and comfortable soft dry in-ear electrode systems could, therefore, have significant implications for both mHealth and human-machine interface (HMI) applications. This research evaluates the quality of the ECG signal obtained with soft dry electrodes inserted in the ear canal. An earplug with six soft dry electrodes distributed around its perimeter was designed for this study, allowing for the analysis of the signal coming from each electrode independently with respect to a common reference placed at different positions on the body of the participants. An analysis of the signals in comparison with a reference signal measured on the upper right chest (RA) and lower left chest (LL) was performed. The results show three typical behaviors for the in-ear electrodes. Some electrodes have a high correlation with the reference signal directly after inserting the earplug, other electrodes need a settling time of typically 1-3 min, and finally, others never have a high correlation. The SoftPulseTM electrodes used in this research have been proven to be perfectly capable of measuring physiological signals, paving the way for their use in mHealth or HMI applications. The use of multiple electrodes distributed in the ear canal has the advantage of allowing a more reliable acquisition by intelligently selecting the signal acquisition locations or allowing a better spatial resolution for certain applications by processing these signals independently.

A skin-conformal and breathable humidity sensor for emotional mode recognition and non-contact human-machine interface

Noncontact humidity sensor overcomes the limitations of its contact sensing counterparts, including mechanical wear and cross infection, which becomes a promising candidate in healthcare and human-machine interface application. However, current humidity sensors still suffer the ubiquitous issue of uncomfortable wear and skin irritation hindering the long-term use. In this study, we report a skin-conformal and breathable humidity sensor assembled by anchoring MXenes-based composite into electrospun elastomer nanofibers coated with a patterned electrode. This composite is highly sensitive to the water molecules due to its large specific surface area and abundant water-absorbing hydroxyl groups, while the elastomeric nanofibers provide an ultrathin, highly flexible, and permeable substrate to support the functional materials and electrodes. This sensor presents not only excellent air permeability (0.078 g cm−2 d−1), high sensitivity (S = 704), and fast response/recovery (0.9 s/0.9 s), but also high skin conformability and biocompatibility. Furthermore, this humidity sensor is confirmed to realize the recognition of motional states and emotional modes, which provides a way for the advanced noncontact human-machine interface.

Synergistic advancements in high-performance flexible capacitive pressure sensors: structural modifications, AI integration, and diverse applications.

The development of flexible pressure sensors for monitoring human motion and physiological signals has attracted extensive scientific research. However, achieving low monitoring limits, a wide detection range, large bending stresses, and excellent mechanical stability simultaneously remains a serious challenge. With the aim of developing a high-performance capacitive pressure sensor (CPS), this paper introduces the successful preparation of a single-walled carbon nanotube (SWNT)/polydimethylsiloxane (S-PDMS) composite dielectric with a foam-like structure (high permittivity and low elasticity modulus) and MXene/SWNT (S-MXene) composite film electrodes with a micro-crumpled structure. The above structurally modified CPS (SMCPS) demonstrated an excellent response output during pressure loading, achieving a wide pressure detection range (up to 700 kPa), a low detection limit (16.55 Pa), fast response/recovery characteristics (48/60 ms), enhanced sensitivity across a wide pressure range, long-term stability under repeated heavy loading and unloading (40 kPa, >2000 cycles), and reliable performance under various temperature and humidity conditions. The SMCPS demonstrated a precise and stable capacitive response in monitoring subtle physiological signals and detecting motion, owing to its unique electrode structure. The flexible device was integrated with an Internet of Things module to create a smart glove system that enables real-time tracking of dynamic gestures. This system demonstrates exceptional performance in gesture recognition and prediction with artificial intelligence analysis, highlighting the potential of the SMCPS in human-machine interface applications.

A Novel Approach to Surface EMG-based Gesture Classification Using a Vision Transformer Integrated with Convolutive Blind Source Separation.

A robust pattern recognition framework is required for ideal real-time human-machine interface (HMI) applications. Convolutional neural networks and recurrent neural networks have been widely used for the classification of gestures based on electromyography (EMG), but few studies have demonstrated the effectiveness of using a vision transformer for this purpose. Additionally, the accuracy achieved is influenced by the efficacy of the preprocessing pipeline. This study assessed ViT with and without an attention mechanism for precise motor intent decoding by investigating various input features and integrating convolutive blind source separation (BSS) preprocessing. All investigations were carried out with two open-access high-density surface EMG datasets of 34 and 21 hand gestures recorded from 20 and 5 healthy subjects respectively. Integration of centering and optimal extension factors resulted in better performance with raw input. However, spatial whitening increased the model's sensitivity to noise. The best-performing BSS-integrated convolution vision transformer model (BSS-CViT) model yielded an accuracy of 96.61% and 91.98% on test datasets one and two. This is a promising result for future studies in real-time HMI applications. The code implementation results reported in this study are available on GitHub. https://github.com/deremustapha/BSS-ViT.