End-to-end motion detection via multi-scale spatial-temporal feature fusion for dual-view 3D macaque behavior quantification

Superelastic and highly sensitive conductive hydrogel sensor enabled by spatially confined assembly of MXene within bacterial cellulose network.

Multi-functional zwitterionic glycerylphosphorylcholine hydrogel for human motion detection and human-machine interaction.

A lightweight multi-scale neural network for leg movement detection during sleep

Ground-based remote sensing of seismic and geophysical displacements remains a major challenge due to environmental hazards, signal attenuation, and practical deployment limitations of traditional seismometers. In this study, we present a detailed design, implementation, and performance evaluation of a Moiré-based apparatus for remote ground displacement measurement. The system operates by detecting fringe shifts formed between a fixed and a displaced grating, with displacement magnified through controlled angular superposition. We systematically assess each component of the system, including telescope optics, imaging sensors, and grating configurations, to optimize spatial resolution, contrast, and robustness under varying environmental conditions. A digital approach for fringe generation was employed, allowing controlled magnification and improved sensitivity without the need for physical alignment of dual gratings. Indoor experiments under low-turbulence conditions validated the system’s capability to detect displacements as small as 50μm. Subsequent outdoor trials at different distances demonstrated successful measurement of both square-wave and seismic-like displacements despite increased atmospheric turbulence and wind. The results confirm the system’s ability to perform real-time, long-range, non-contact displacement monitoring with high accuracy and resilience to environmental variability. This study establishes a foundation for the application of Moiré-based sensing in challenging field conditions, including volcanic and seismic zones.

Abstract—Recent advancements in Human–Computer Inter-action (HCI) have focused on developing multimodal systems that enable intuitive and contactless communication between users and machines. Traditional input devices such as keyboards and mice restrict accessibility and limit natural interaction, prompting research into gesture and voice-based interfaces. Numerous studies have explored vision-based gesture recognition using machine learning and computer vision frameworks like MediaPipe and OpenCV, enabling real-time detection of hand movements for cursor control, clicking, and scrolling actions. Similarly, speech recognition technologies leveraging deep learn-ing have evolved to convert spoken language into digital text with high accuracy, facilitating command execution and dictation. Integrating these two modalities—gesture navigation and voice recognition—enhances system adaptability and usability, partic-ularly for users with physical impairments or in touch-restricted environments. This survey indicates a growing shift toward hybrid interfaces that utilize built-in sensors such as cameras and microphones to achieve efficient, low-cost, and cross-platform operation. This convergence of gesture and speech modalities forms the foundation for developing real-time multimodal HCI systems capable of providing seamless, hands-free interaction without external hardware dependencies. Keywords—Human–Computer Interaction (HCI), Gesture Recognition, Voice Recognition, Touchless Interaction

The incorporation of naturally derived bioactive compounds (BAC) into biodegradable polymer matrices presents a promising and sustainable approach to enhancing the performance of triboelectric nanogenerators (TENGs) for next-generation self-powered electronic devices. In this study, poly(vinyl alcohol) (PVA) was added with bioactive constituents extracted from turmeric (Curcuma longa - CL), garlic (Allium sativum - AS), and ginger (Zingiber officinale - ZO) through a solution casting technique to fabricate BAC@PVA composites. These plant-based additives, rich in functional groups and phytochemicals, significantly improved the triboelectric properties by enhancing surface roughness, dielectric behavior, and interfacial charge transfer. The structural, morphological, elemental, and chemical characteristics of the composites are thoroughly examined using XRD, SEM, EDS, and FTIR analyses. TENG devices are fabricated using the BAC@PVA composites as tribopositive layers and PVDF as the tribonegative counterpart. Among the fabricated devices, the CL@PVA-TENG demonstrated superior electrical output, achieving a peak voltage of 302 V and current of 62 μA, marking a notable improvement over pristine PVA-based TENGs. The harvested energy successfully powers capacitors and 57 blue LEDs, demonstrating practical viability. Additionally, the device functions as a self-powered sensor, capable of detecting a wide range of gestures and human interactions with high sensitivity and reliability. These multifunctional capabilities enable practical applications in wearable healthcare monitoring, interactive electronics, and smart human-machine interfaces, where precise detection of motion and touch provides real-time, user-centered benefits. Overall, this study demonstrates that reinforcing PVA with bioactive compounds offers an eco-friendly and efficient strategy for developing sustainable energy-harvesting and self-powered sensing devices.

Street lighting is essential for nighttime traffic safety and public security, yet many conventional installations still operate at constant full brightness from dusk to dawn regardless of road activity. This practice causes unnecessary energy consumption and increases operational costs, particularly in low-traffic periods. This study aimed to design and evaluated a low-cost intelligent street lighting prototype that combines a passive infrared (PIR) motion sensor and a light dependent resistor (LDR) ambient light sensor to reduce full-brightness operating time while maintaining responsive illumination at night. An Arduino-based controller was implemented using a state-based control strategy with three operating modes: OFF during sufficient daylight, DIM standby at night when no motion was detected, and BRIGHT mode when motion was detected, followed by a configurable hold time before returning to DIM. The prototype was tested under four scenarios representing daylight, nighttime idle, nighttime motion, and motion stop conditions, with repeated trials and serial logging of sensor readings, state transitions, and pulse width modulation output levels. The results showed reliable state behavior across scenarios, rapid activation from motion detection to BRIGHT mode with a mean response time of 0.42 s, and consistent hold-time performance near the 30 s target. During a 30-minute nighttime mixed-activity test, the system operated in DIM mode for 62% of the time and in BRIGHT mode for 38%, yielding an estimated 43.4% relative energy reduction compared with an always-on full-brightness baseline. The findings indicate that integrating PIR motion sensing and LDR-based ambient gating provides a practical and replicable pathway to improve street lighting energy efficiency without sacrificing on-demand illumination for road users.

Neural vision restoration is a rapidly advancing discipline at the intersection of neuroscience, bioengineering, and ophthalmology. This review synthesizes emerging strategies to restore visual perception through retinal prostheses, optic nerve and thalamic implants, cortical brain-computer interfaces (BCIs), optogenetics, and non-invasive stimulation. Although initial experiments have demonstrated primitive visual abilities such as light perception and motion detection, artificial vision remains cognitively demanding and fundamentally different from natural vision. Advances in artificial intelligence and machine learning may enable adaptive, closed-loop systems that optimize stimulation, enhance low-light vision, and integrate environmental inputs for more intelligible percepts. At the same time, a growing understanding of neural plasticity, cortical remapping, and perceptual learning highlights the need for multidisciplinary strategies in visualrehabilitation. Ethical and regulatory concerns, including informed consent, data protection, neural enhancement, and equitable access, remain central to responsible implementation. The potential of BCIs to bypass the eye entirely, and of neuroprosthetics to be used in spaceflight, disaster response, or military medicine, expands the applications of vision restoration beyond blindness alone. Bridging technological, clinical, and ethical strategies in this review outlines the challenges and opportunities that define the future of neural ophthalmology. Ultimately, restoring sight will require not only functioning hardware, but systems compatible with the reorganized brain and the livedexperience of visual loss.

Detecting and tracking object has become a critical aspect in modern automation, security setup, and robotic technologies, as these systems modifies efficiency and safety, where fast and accurate detection is critical. we present a hybrid radar detection system that combines multiple sensing technologies to results were accurate detection. It detected object within a 50 cm range. The system setup an HB100 Doppler radar module for motion and speed detection with an HC-SR04 ultrasonic sensor for accurate distance measurement. A micro-servo motor sweep 180° scanning, while an ESP32 microcontroller manages sensor data, real-time computation, visualization, and alert generation. An OLED display provides a live sweep showing the object’s distance, speed, and angular position, and a camera module provides video feed. This study shows research contributions from the past decade to identify advancement in hybrid radar-based detection systems and their application ultrasonic radar, Doppler sensor, and IoT-enabled alert models. To improve accuracy, the system includes local alerts through LEDs, a buzzer, and a speaker’s, along with remote notifications via an SMS module. An RTC module ensures accurate time stamping of events, and an actuation mechanism activates when a risk is detected. The prototype shows stable performance, with correct object tracking and reduced false alarms due to the used of several sensor calculating distance, speed, and angle information, this model enhances situational awareness.

To discern speech or appreciate music, the human auditory system detects how pitch changes over time (pitch motion). Here, using psychophysics, computational modelling, functional neuroimaging and analysis of recorded speech, we ask whether humans can detect pitch motion using computations analogous to those used by the visual system. We adapted stimuli from studies of vision to create novel auditory correlated noise stimuli that elicited robust pitch motion percepts. In psychophysical experiments, we discovered that humans can judge pitch direction from spectrotemporal intensity correlations. Robust sensitivity to negative spectrotemporal correlations is a direct analogue of illusory 'reverse-phi' motion in vision, constituting a new auditory illusion. Functional MRI measurements in auditory cortex supported the hypothesis that human auditory processing may employ pitch direction opponency. Linking lab findings to real-world perception, we analysed recordings of English and Mandarin speech and found that pitch direction was signalled by both positive and negative spectrotemporal correlations, suggesting that sensitivity to both types confers ecological benefits. This work reveals how motion detection algorithms sensitive to local correlations are deployed by the central nervous system across disparate modalities (vision and audition) and dimensions (space and frequency).

Fluorinated bi-layer E-textile enabled by electrospinning rotary collection for self-powered dual-functional electrochemical sweat sensing and motion detection

In contemporary security management, the integration of automated deterrents and real-time notification systems is essential for effective property protection. This study presents the design and implementation of a Proximity-Activated Guard Dog Sound System with an Embedded GSM Module, an intelligent security solution that combines psychological deterrence with remote alerting capabilities. The system utilizes an Arduino microcontroller as the central processing unit, interfaced with a Passive Infrared (PIR) sensor for high-sensitivity motion detection. Upon detecting an unauthorized heat signature, the system executes a dual-action response: it triggers a DFPlayer Mini audio module to broadcast high-decibel, realistic dog barking sounds through an external speaker, and simultaneously utilizes a SIM800L GSM module to transmit an instant SMS alert to the property owner’s mobile device. Experimental results indicate that the system effectively differentiates between static environments and human-scale movement, providing a cost-effective alternative to maintaining live guard animals or expensive CCTV monitoring services. The inclusion of the GSM layer ensures that security breaches are communicated within seconds, regardless of the user's geographical location. This project demonstrates a scalable model for smart home security, offering a proactive approach to crime prevention through acoustic simulation and cellular communication.

Decoupled and high-sensitivity strain-temperature bimodal sensors based on chitosan-enhanced hydrogel with local strain concentration.

Realizing high-performance wood-based piezoresistive sensing through an interfacial bridging approach for wearable functional integration.

Self-healing repairing flexible pressure sensor based on liquid metal-graphene aerogel/multi-walled carbon nanotubes-polyurethane composites film for human motion detection

Trehalose-toughened transparent hydrogel with fast-healing and anti-freezing capability for precision human motion detection

Infrared (IR) machine vision systems have advanced significantly in a range of applications, including autonomous driving, security monitoring, and intelligent night vision. Despite advances in optoelectronic memristors by mixed optical and electrical operation, achieving infrared-specific fully light-modulated vision systems remains challenging. Here, we demonstrate a plasmonic optoelectronic memristor based on Te nanowires-Au nanoparticles/ι-carrageenan film. Localized surface plasmon resonanceassisted optical excitation endows the device with non-volatile, IR-programmable conductance states that can be selectively erased by visible light, yielding an all-photonic write/erase scheme without electrical intervention. Exploiting this reversible photonic plasticity, we construct an IR vision system capable of in-sensor Boolean logic and motion detection under complete darkness. An optical neural network trained on the reversible conductance dynamics attains 91.4% recognition accuracy for moving objects. This work proposes a fully light-modulated optoelectronic memristor that may promote the future development of efficient IR machine vision systems.

ABSTRACT Flexible pressure sensors play an increasingly important role in wearable electronics, particularly in areas such as healthcare monitoring, human–machine interaction, and motion detection. Despite significant progress, maintaining high sensitivity while ensuring stable performance over a broad pressure range remains a major challenge. To address this, the present work develops a flexible capacitance pressure sensor using a PVDF‐GO‐ZnO composite as the dielectric layer. Zinc oxide nanoparticles were synthesized through a solvothermal reflux method, and graphene oxide was prepared using the Modified Hummers' process. The composite layer was formed by spin‐coating PVDF incorporated with GO and ZnO, enabling a systematic evaluation of how these fillers influence the dielectric properties and sensing behavior. The resulting sensor shows excellent performance in the low‐pressure range (0–3 kPa), achieving a sensitivity of 2.73655 kPa −1 , response and recovery times (0.5 and 0.48 s), hysteresis (3.53%), and stable operation over 3500 s. These results highlight the device's robustness and suitability for wearable applications, particularly in health diagnostics, motion tracking, and speech‐based monitoring.

In pursuit of high-performance flexible strain sensors, achieving an optimal trade-off among linearity, sensitivity, and strain sensing range remains a critical challenge. Inspired by the wrinkled-leaf viburnum, we develop a Janus sensor that replicates its asymmetric structure. It comprises a dense, micro-wrinkled natural rubber (NR)/graphene (GRs) top layer and a loose NR/carbon nanotubes (CNTs) bottom layer, fabricated via facile layer-by-layer filtration and pre-stretching strategy. This bio-inspired design enables the sensor with a synergistic sensing mechanism: wrinkle-guided microcrack ensures highly sensitive linear response at low strains; strain-phase division maintains signal continuity at medium strains; and parallel conductive circuits provide robustness at high strains. As a result, the sensor achieves an exceptional combination of ultra-high linearity (R2 > 0.999) and sensitivity (gauge factors, GF > 14) across 0-100% strain, with a wide sensing range (> 400%) and fast response (0.16 s). We demonstrate its practical value in human motion detection, physiological signal monitoring, and an intelligent glove system for gesture recognition and human-machine interaction, highlighting its promising potential for advanced wearable devices and human-machine interactive systems.