A soft closed-loop glove with gesture sensing and haptic feedback for real-time human-machine interfaces

Tire deformation monitoring is a critical requirement for improving vehicle safety, performance, and intelligent transportation systems. However, most existing flexible strain sensors either lack directional sensitivity or have not been validated in real-world driving environments, limiting their practical application in smart tires. In this work, we report the fabrication of a flexible piezoresistive strain sensor based on a porous laser-induced graphene (LIG) network embedded in an Ecoflex elastomer matrix, with integrated directional force recognition. The LIG–Ecoflex sensor exhibits a high gauge factor of 9.7, fast response and recovery times, and stable performance over 10,000 cycles. More importantly, the anisotropic structure of the LIG enables accurate multi-directional stress recognition when combined with a convolutional neural network (CNN), achieving an overall classification accuracy exceeding 98%. To further validate real-world applicability, the sensor was mounted inside passenger car tires and tested under different loads and speeds. The results demonstrate reliable monitoring of tire deformation with clear correlations to load and velocity, confirming robustness under dynamic driving conditions. This study provides a new pathway for the integration of direction-aware, high-performance strain sensors into intelligent tire systems, with broader potential for wearable electronics, vehicle health monitoring, and next-generation Internet of Vehicles applications.

Flexible sensing technology offers significant advantages in gesture recognition and human–machine interactions. However, existent smart gloves based on flexible sensors still have limitations in their security mechanisms; thus, they are unsuitable for high‐risk environments where identity verification and system protection are critical. This study proposes an innovative smart glove system in which person recognition and gesture control functionalities are integrated. The glove utilizes a triboelectric nanogenerator based on a polyvinyl‐alcohol–sodium‐alginate–conductive‐polyaniline (PSP) hydrogel electrode as the sensor and exhibits high sensitivity, fast response/recovery, and fatigue resistance. These properties are primarily attributed to the excellent stretchability, conductivity, and self‐healing ability of the PSP hydrogel. Because of the coordinated design of the hardware–software architecture, the smart glove enables precise control of drone flight postures via contact sensing. The glove also leverages non‐contact sensing to recognize personalized fingertip trajectories, enabling user recognition and unlocking the aforementioned drone control interface. This design improves not only the flexibility and portability of drone operation but also safety of weapon systems in future battlefield environments.

Flexible and wearable piezoelectric sensors have gained attention for applications in human‐machine interfacing (HMI) and the artificial intelligence of things (AIoT). In this study, antimony‐doped barium titanate (Ba 0.3 Sb₀.₇TiO 3 ) was used to fabricate a piezoelectric nanogenerator (PENG) for finger movement sensing and gesture recognition. The polymer‐to‐particle ratio was optimized, with 20 wt.% yielding the best performance. The optimized PENG generated a maximum output of 50 V and 1.5 µA under applied force, validated by charging commercial capacitors and powering LEDs. For real‐time applications, the device was scaled to finger size and integrated into a wearable glove capable of detecting finger motion. The generated electrical signals were processed using convolutional neural networks (CNNs), converting the signals into readable text through deep learning. Using this approach, the glove successfully recognized the words “SOS” and “HELLO,” demonstrating the potential of the PENG for smart wearable applications. This work highlights the integration of piezoelectric sensing with AI‐enabled gesture recognition, offering a promising route for advanced wearable healthcare devices and interactive technologies.

In intelligent manufacturing and robotic technology, various sensors must be integrated with equipment. In addition to traditional sensors, stretchable sensors are particularly attractive for applications in robotics and wearable devices. In this study, a piezoresistive stretchable strain sensor based on laser-induced graphene (LIG) was proposed and developed. A three-dimensional, porous LIG structure fabricated from polyimide (PI) film using laser scanning was used as the sensing layer of the strain sensor. Two LIG pattern structures (parallel and vertical) were fabricated and integrated within the LIG strain sensors. Scanning electron microscopy, an X-ray energy dispersive spectrometer, and Raman scattering spectroscopy were used to examine the microstructure of the LIG sensing layer. The performance and strain sensing properties of the parallel and vertical stretchable LIG strain sensors were investigated in tensile tests. The relative resistance changes and the gauge factors of the parallel and vertical LIG strain sensors were quantified. The parallel strain sensor achieved a high gauge factor of 15.79 in the applied strain range of 10% to 20%. It also had high sensitivity, excellent repeatability, good durability, and fast response times during the tensile experiments. The developed LIG strain sensor can be used for the real-time monitoring of human motions such like finger bending, wrist bending, and throat swallowing.

Stretchable strain sensors that can detect and monitor human motion have attracted enormous attention in the field. To realize fully the translation of materials for stretchable strain sensors, it is important to make the fabrication way less complex and lessen the production cost. This paper investigated low-cost graphite-VHB strain sensors prepared by simple friction-transfer method. Designed by friction-transfer method, the graphite-VHB strain sensor possesses good flexibility, light weight, high stretchability and tunable range of GF. The gauge factor of the sample was 20 when strain rate was 3% and could be over 1000 when the strain rate was 50%. The addition of protection layer and pre-aging process have contributions to expand the GF of the sample. By attaching our strain sensor to the sample of mechanical arm, the resistance change of the sensor perfectly fitted to the deformation of the arm, which shows its potential in application of soft robotics and mechanical fields. The results showed the friction-transfer method is a simple and low-cost method, featured with environmental advantage that avoids chemical wastes.

Stretchable strain sensors play a vital role in monitoring human body movements and have been widely applied in human–machine interface (HMI) systems. Conventional planarfabricated sensors attached to curved surfaces often suffer from low fabrication efficiency and reduced measurement accuracy due to poor conformability. This study presents a laser direct writing–based approach for the direct fabrication of double‐layer strain sensor arrays, interconnects and electrodes on complex curved, stretchable substrates. The resulting strain sensor exhibits a high gauge factor of 10⁶ at 80% strain and a minimum detectable strain as low as 0.1%, enabling precise monitoring of joint motion across a bending range of 0° to 90°. Additionally, tuning the material composition and structural design enables precise control over the sensor's electromechanical responses to suit various HMI applications. As a demonstration, an HMI with integrated strain sensors and interconnect arrays is directly fabricated onto the 3D surfaces of a glove. This system enables the robotic hand to accurately mimic multiple human gestures, highlighting its potential for advanced human–machine interaction.

Flexible and stretchable strain sensors have attracted growing interest for applications in health monitoring and human-machine interfaces. Conventional crack-based resistive strain sensors can achieve high sensitivity and large strain ranges by exploiting the opening and closing of randomly distributed microcracks. However, the uncontrolled evolution of these cracks during prolonged use often leads to performance degradation and poor reliability. In this work, we present a novel resistive strain sensor featuring predefined hourglass-shaped cracks that precisely regulate resistance variations through controlled crack opening and closure under stretching and release. A serpentine array of hourglass-shaped cracks is laser-cut into a 3D-printed conductive TPU film, after which the crack gaps are closed by releasing a pre-stretched elastomer substrate. Through geometric design, crack contact and evolution are precisely controlled, enabling the cracks to open stably and symmetrically under tensile strain, thereby producing predictable and linear resistance changes. This design enables high sensitivity over a broad linear operating range. The sensor also exhibits excellent recoverability after overstretching well beyond its working window and shows negligible resistance variation under bending, highlighting its robustness for wearable applications. To demonstrate its practical utility, the sensor was integrated into athletes' strength-training tasks, and the acquired signals were analyzed using machine learning algorithms to recognize motion patterns and provide early warnings of over-exercise. These capabilities support real-time load adjustment and fatigue-risk management, underscoring the sensor's strong potential for wearable health monitoring and human-machine interface applications.

Organisms rely on seamlessly integrated sensing and action feedback loops to dynamically adjust their behavior for adaptation, learning, and survival. Biomimetic self‐sensing hydrogel soft robots are an emerging class of intelligent systems that integrate actuation and sensing functions within a single compliant structure. They are characterized by tissue‐like softness, high adaptability, and the ability to monitor deformation and environmental interactions in real time. Such intelligent systems have great potentials in soft robotics, biomedical devices, and human–machine interaction. This review comprehensively summarizes recent advances in biomimetic self‐sensing hydrogel soft robots. It covers biomimetic microstructured hydrogel actuators responsive to diverse external stimuli, as well as biomimetic microstructured hydrogel sensors for detecting pressure, strain, and multidimensional signals. Endowing hydrogel sensors with stimulus‐responsive actuation or imparting hydrogel actuators with sensing capabilities allows for the integration of both sensing and actuation functions into a single hydrogel device, thereby resulting in a self‐sensing hydrogel actuator. The biomimetic self‐sensing with microstructured hydrogel actuators driven by thermal, optical, magnetic, and tactile stimuli is described in detail. Some in‐depth analysis of current challenges and future perspectives is provided at the end of the review. This review offers valuable insights into the development of autonomous, intelligent soft robotic systems.

Gesture recognition based on surface electromyography (sEMG) has been widely used in the field of human-machine interaction (HMI). However, sEMG has limitations, such as low signal-to-noise ratio and insensitivity to fine finger movements, so we consider adding A-mode ultrasound (AUS) to enhance the recognition impact. To explore the influence of multisource sensing data on gesture recognition and better integrate the features of different modules. We proposed a multimodal multilevel converged attention network (MMCANet) model for multisource signals composed of sEMG and AUS. The proposed model extracts the hidden features of the AUS signal with a convolutional neural network (CNN). Meanwhile, a CNN-LSTM (long-short memory network) hybrid structure extracts some spatial-temporal features from the sEMG signal. Then, two types of CNN features from AUS and sEMG are spliced and transmitted to a transformer encoder to fuse the information and interact with sEMG features to produce hybrid features. Finally, the classification results are output employing fully connected layers. Attention mechanisms are used to adjust the weights of feature channels. We compared MMCANet's feature extraction and classification performance with that of manually extracted sEMG-AUS features using four traditional machine-learning (ML) algorithms. The recognition accuracy increased by at least 5.15%. In addition, we tried deep learning (DL) methods with CNN on single modals. The experimental results showed that the proposed model improved 14.31% and 3.80% over the CNN method with single sEMG and AUS, respectively. Compared with some state-of-the-art fusion techniques, our method also achieved better results.

Stretchable strain sensors with well-controlled sensitivity and stretchability are crucial for applications ranging from large deformation monitoring to subtle vibration detection. Here, based on single-metal material on the elastomer and one-pot evaporation fabrication method, we realize controlled strain sensor performance via a novel programable cracking technology. Specifically, through elastomeric substrate surface chemistry modification, the microcrack generation and morphology evolution of the strain sensing layer is controlled. This process allows for fine tunability of the cracked film morphology, resulting in strain sensing devices with a sensitivity gauge factor of over 10 000 and stretchability up to 100%. Devices with a frequency response up to 5.2 Hz and stability higher than 1000 cycles are reported. The reported strain sensors, tracking both subtle and drastic mechanical deformations, are demonstrated in healthcare devices, human-machine interaction, and smart-home applications.

In recent years, gesture recognition and speech recognition, as important input methods in Human–Computer Interaction (HCI), have been widely used in the field of virtual reality. In particular, with the rapid development of deep learning, artificial intelligence, and other computer technologies, gesture recognition and speech recognition have achieved breakthrough research progress. The search platform used in this work is mainly the Google Academic and literature database Web of Science. According to the keywords related to HCI and deep learning, such as “intelligent HCI”, “speech recognition”, “gesture recognition”, and “natural language processing”, nearly 1000 studies were selected. Then, nearly 500 studies of research methods were selected and 100 studies were finally selected as the research content of this work after five years (2019–2022) of year screening. First, the current situation of the HCI intelligent system is analyzed, the realization of gesture interaction and voice interaction in HCI is summarized, and the advantages brought by deep learning are selected for research. Then, the core concepts of gesture interaction are introduced and the progress of gesture recognition and speech recognition interaction is analyzed. Furthermore, the representative applications of gesture recognition and speech recognition interaction are described. Finally, the current HCI in the direction of natural language processing is investigated. The results show that the combination of intelligent HCI and deep learning is deeply applied in gesture recognition, speech recognition, emotion recognition, and intelligent robot direction. A wide variety of recognition methods were proposed in related research fields and verified by experiments. Compared with interactive methods without deep learning, high recognition accuracy was achieved. In Human–Machine Interfaces (HMIs) with voice support, context plays an important role in improving user interfaces. Whether it is voice search, mobile communication, or children’s speech recognition, HCI combined with deep learning can maintain better robustness. The combination of convolutional neural networks and long short-term memory networks can greatly improve the accuracy and precision of action recognition. Therefore, in the future, the application field of HCI will involve more industries and greater prospects are expected.

Stretchable strain sensors with dentate groove structure for enhanced sensing recoverability

With the rapid development of soft electronics, flexible and stretchable strain sensors are highly desirable. However, coupling of high sensitivity and stretchability in a single strain sensor remains a challenge. Herein, a kind of conductive elastomer is constructed with poly(dimethylsiloxane) (PDMS) and silylated cellulose nanocrystal (SCNC)/carbon nanotube (CNT) nanohybrids through a facile one-pot solution-casting method. The hydrophobic SCNCs can effectively facilitate the dispersion of CNTs in PDMS and synergistically improve the interfacial compatibility between CNTs and the PDMS matrix, resulting in favorable stress and electron transfer in the polymer network. Due to the outstanding electrical conductivity of CNTs and the excellent dispersity and high mechanical performance of SCNCs, combined with the good compatibility between SCNC-mediated carbon nanotubes (SCNC-CNTs) and PDMS, the resulting composite elastomer (SCNC-CNT/PDMS) shows high electrical conductivity (∼2.77 S m-1), tensile strength (∼5.72 MPa), and fatigue resistance properties. The strain sensor assembled by SCNC-CNT/PDMS demonstrates a high strain range above 100%, appealing strain sensitivity with a gauge factor of 37.11 at 50-100% strain, and long-term stability and durability, which is capable of monitoring both real-time human motions and acoustic vibrations. This work paves a new way for the design and controllable preparation of flexible and stretchable conductive elastomers, demonstrating promising applications in wearable devices and intelligent electronics.

In-situ enhanced anchor-free deep CNN framework for a high-speed human-machine interaction

An adhesive, highly stretchable and low-hysteresis alginate-based conductive hydrogel strain sensing system for motion capture