Search

This work introduces a silent speech interface (SSI), proposing a few-layer graphene (FLG) strain sensing mechanism based on thorough cracks and AI-based self-adaptation capabilities that overcome the limitations of state-of-the-art technologies by simultaneously achieving high accuracy, high computational efficiency, and fast decoding speed while maintaining excellent user comfort. We demonstrate its application in a biocompatible textile-integrated ultrasensitive strain sensor embedded into a smart choker, which conforms to the user's throat. Thanks to the structure of ordered through cracks in the graphene-coated textile, the proposed strain gauge achieves a gauge factor of 317 with <5% strain, corresponding to a 420% improvement over existing textile strain sensors fabricated by printing and coating technologies reported to date. Its high sensitivity allows it to capture subtle throat movements, simplifying signal processing and enabling the use of a computationally efficient neural network. The resulting neural network, based on a one-dimensional convolutional model, reduces computational load by 90% while maintaining a remarkable 95.25% accuracy in speech decoding. The synergy in sensor design and neural network optimization offers a promising solution for practical, wearable SSI systems, paving the way for seamless, natural silent communication in diverse settings. [ABSTRACT FROM AUTHOR]

With the rapid development of Internet of Things (IoT) technologies, traditional disease diagnoses carried out in medical institutions can now be performed remotely at home or even ambient environments, yielding the concept of the Internet of Health Things (IoHT). Among the diverse IoHT applications, inertial measurement unit (IMU)-based systems play a significant role in the detection of diseases in many fields, such as neurological, musculoskeletal, and mental. However, traditional numerical interpretation methods have proven to be challenging to provide satisfying detection accuracies owing to the low quality of raw data, especially under strong electromagnetic interference (EMI). To address this issue, in recent years, machine learning (ML)-based techniques have been proposed to smartly map IMU-captured data on disease detection and progress. After a decade of development, the combination of IMUs and ML algorithms for assistive disease diagnosis has become a hot topic, with an increasing number of studies reported yearly. A systematic search was conducted in four databases covering the aforementioned topic for articles published in the past six years. Eighty-one articles were included and discussed concerning two aspects: different ML techniques and application scenarios. This review yielded the conclusion that, with the help of ML technology, IMUs can serve as a crucial element in disease diagnosis, severity assessment, characteristic estimation, and monitoring during the rehabilitation process. Furthermore, it summarizes the state-of-the-art, analyzes challenges, and provides foreseeable future trends for developing IMU-ML systems for IoHT. [ABSTRACT FROM AUTHOR]

Land allocation and pricing substantially impact carbon emissions, yet their nonlinear effects remain understudied. This research employs ensemble machine learning models to examine the complex relationships between land conveyance and per capita carbon emissions across 104 major Chinese cities from 2009 to 2017. The results reveal that keeping industrial land allocations below 35% helps reduce emissions, whereas higher ratios increase emissions. Allocating over 8% and 33% to business and public land respectively also lowers emissions. Land prices demonstrate heterogeneity – a higher residential land price promotes efficiency only when its relative price level to the comprehensive land price is below 1.1. The findings highlight customised policies balancing development and emissions reduction, based on local conditions and development stages, can forge sustainable pathways. Overall, the nonlinear modelling quantifies nuanced emissions responses to land allocation thresholds and strategic pricing incentives. By considering these complex mechanisms, urban planners can devise tailored strategies that simultaneously nurture growth and curb emissions. The novel method and evidence-based insights contribute to planning support systems and sustainable policy-making. • Estimating land conveyance's nonlinear effects on emissions by ensemble learning. • Allocation and price of land display nonlinear relationships on carbon emissions. • Land use within suitable ratios and pricing incentives promotes carbon reductions. [ABSTRACT FROM AUTHOR]

With the rapid development of touch sensing technology, keystroke authentication has been proved to be a more secure and reliable identification method. However, emotion has a big influence on users’ keystroke habits, a condition not considered in previous studies, yet greatly affect the accuracy of keystroke authentication. In this article, we propose a new strategy for keystroke authentication applicable to multiple emotional states based on piezoelectric touch panel, which provides precise force data and enables a higher detection accuracy. In our experiment, keystroke samples composed by force and time features of legitimate user and intruders in different emotions were collected by a piezoelectric touch screen. Four typical machine learning algorithms were used for authentication. Finally, the result was demonstrated, that Random Forest classifier achieved an accuracy of 96.40%, a FAR of 1.02% and a FRR of 8.82%, which is significantly better than the result when only emotion-stable keystroke samples were used to train classifiers. The proposed strategy improves the practicability of keystroke authentication technique in real world application. [ABSTRACT FROM AUTHOR]

Touch orientation detection is important for piezoelectric touch panels to stabilize force-voltage responsivities. Current touch orientation estimation techniques utilize machine learning algorithms for orientation classification. However, environmental noise could weaken the data quality which will result in a lowered detection accuracy. To address this issue, in this article, we present a noise robustness technique, in which different levels of noise data are injected into the training data. The performance of “dirty” data trained model exhibits a good performance (average mean absolute error (MAE) of 7.8 degrees) among signal-to-noise ratio (SNR) from 3 dB to 40 dB, indicating that an improved user experience can be obtained. [ABSTRACT FROM AUTHOR]

Force touch function in interactive displays gains increasing interests due to its ability in detecting 3D information from customers, giving rise to plenty of novel and advanced services for users and therefore increasing the efficiency of human‐machine interactivity. In this article, piezoelectric based force touch architecture and machine learning algorithms are smoothly combined for implementation of user‐centered force level recognition and user identity authentication. [ABSTRACT FROM AUTHOR]

With the rapid development of electric and information technologies, security issues are becoming a pressing concern. This has given rise to a plethora of authentication techniques, leading among them being the fixed-password method due to its simple mechanism and independence from any specific hardware in the modern smartphone. While fixed-password methods enjoy benefits such as ease of use, potential security issues associated with password leakage cannot be ignored. This article, investigates an alternative strategy based on user keystrokes. Here, the user touch times and force features are extracted from a piezoelectric force touch panel which is an integral part of the hardware. Three broadly adopted machine learning classifiers are used for the collected data set, finally achieving an Equal Error Rate (EER) of 0.720%. [ABSTRACT FROM AUTHOR]

Accurate snow depth (SD) information at global scales is highly important for validation and calibration of hydrological model, snowmelt runoff prediction, and global water supplies estimation. Passive microwave remote sensing has been proven to be an efficient method to detect SD at regional or global scales. However, numerous heterogeneous snow property parameters and forest parameters affect the microwave radiation characteristics, which is a great challenge for SD retrieval. In this study, a novel SD retrieval algorithm that comprehensively considered the multiple sensitive factors was developed at global scales by coupling the Microwave Emission Model of Layered Snowpacks (MEMLS) model and machine learning (ML) technique. The exponential correlation length (Pec), snow density (ρ), and forest cover fraction (f f o r e s t) are three sensitive parameters for retrieving SD from passive microwave observations based on the extended Fourier amplitude sensitivity test (EFAST) algorithm. The key parameter Pec , which characterizes the snow microstructure information inside snowpacks, was optimized by minimizing the brightness temperature between satellite observations and microwave radiative transfer model simulations by integrating the MEMLS, forest impact, and atmospheric radiation attenuation. The random forest (RF) ML technique was used, and the optimal combination of independent variables was selected by correlation heatmap and importance ranking of variables. Six selected essential parameters, including Pec , vertical polarized brightness temperature differences between 18.7 and 36.5 GHz, ρ from ERA5, f f o r e s t , elevation, and longitude together with target SD output were used to train the RF model. Seventy percent of the entire data at meteorological stations were randomly chosen to train the RF model, and the remaining 30% were adopted as the validation data from 2013 to 2020 to ensure the mutual independence of training data and validation data. Compared with independent in situ measurements, the overall RMSE of SD is 13.6 cm, which was noticeably lower than those of the international SD remote sensing products and reanalysis datasets, including AMSR2 (36.2 cm), GlobSnow 3.0 (25.3 cm), ERA5 (38.2 cm), and ERA5-land (29.9 cm). The results demonstrated that the new proposed algorithm, by coupling the MEMLS model with the ML technique, could significantly enhance the overall performance in SD retrieval at global scales. • A novel method to retrieve snow depth (SD) at global scales is proposed by considering snow metamorphism and forest influence. • Exponential correlation length (Pec), snow density, and forest cover fraction are three sensitive parameters for retrieving SD from passive microwave observations. [ABSTRACT FROM AUTHOR]

• Calculating municipal-level carbon emissions using provincial-level patterns have produced biased results. • Time-series smoothing can greatly improve the accuracy of carbon emission estimation. • Aggregation of carbon emissions in China has further increased. • Per-capita carbon emissions of Chinese cities are gradually peaking. Carbon emissions reduction has become a world consensus. Cities have an essential role to play in addressing emission reductions. However, previous studies have estimated China's municipal-level carbon emissions based on provincial-level emission patterns, and such a top-down carbon emission accounting approach has led to biased results. Therefore, this study employed an incremental learning ensemble model and a Savitzky-Golay algorithm tomeasure carbon emission distribution and unearth patterns at the municipal level based on nighttime light (NTL) and regional development characteristics (GDP, population, patents, industry structure). The performance of the proposed method is substantially better than its counterparts in terms of municipal-level estimation (R-square boosted by 20.64%). This research shows significant difference in carbon emission mechanisms between provinces and cities and demonstrates that carbon emissions are time-continuous. It also shows that per-capita carbon emissions are peaking in many China's cities, except in some heavy industrial cities. Our approach provides accurate and dynamic monitoring of municipal emissions in China. [ABSTRACT FROM AUTHOR]

High detection accuracy in piezoelectric-based force sensing in interactive displays has gained global attention. To achieve this, artificial neural networks (ANN)—successful and widely used machine learning algorithms—have been demonstrated to be potentially powerful tools, providing acceptable location detection accuracy of 95.2% and force level recognition of 93.3% in a previous study. While these values might be acceptable for conventional operations, e.g., opening a folder, they must be boosted for applications where intensive operations are performed. Furthermore, the relatively high computational cost reported prevents the popularity of ANN-based techniques in conventional artificial intelligence (AI) chip-free end-terminals. In this article, an ANN is designed and optimized for piezoelectric-based touch panels in interactive displays for the first time. The presented technique experimentally allows a conventional smart device to work smoothly with a high detection accuracy of above 97% for both location and force level detection with a low computational cost, thereby advancing the user experience, and serviced by piezoelectric-based touch interfaces in displays. [ABSTRACT FROM AUTHOR]

As the world's largest developing nation, China has emerged as the predominant carbon emitter due to its swift industrialization and urbanization. Achieving low-carbon development in China necessitates a tailored approach that accounts for each city's unique industrial phase, technological advancements, urbanization level, and resources. This research endeavors to elucidate the nonlinear mechanism influencing carbon emissions in Chinese cities using a machine-learning framework, drawing on panel data from 290 cities from 2000 to 2017. The result indicates that China is experiencing a transition in the primary drivers of carbon emissions from the agriculture sector to the industrial and service sectors. A city transitions to the carbon decoupling phase when the outputs of primary, secondary, and tertiary industries surpass 1500, 125,000, and 100,000 CNY per capita, respectively. Additionally, we observed that expanding the urban built-up area beyond 10% can significantly mitigate carbon emission intensity. However, the study also highlights a critical challenge: the initial emission-reducing effects of innovation in burgeoning high-tech industries are negated by the higher emissions from traditional industries. Moreover, our analysis indicates that R&D investments exceeding RMB 8000 per capita may paradoxically lead to an increase in emissions. With these conclusions of nonlinearity, we emphasize designing policies tailored to the specificities of each city, stressing the importance of adaptability in policy creation. • The carbon emission of the secondary sector is decoupled as output rises. • High tertiary sector output per capita will reduce carbon emissions. • Technological application limits the carbon-reducing effect of innovation output. • Traditional industries mask the low-emission from technological innovation. • Constructing accurate nonlinear driving patterns by ensemble learning algorithms. [ABSTRACT FROM AUTHOR]