Wearable sensors have evolved from niche fitness trackers into a cornerstone of modern biomedical engineering, enabling continuous, non-invasive monitoring of physiological and biochemical markers. Recent breakthroughs in materials science, microfabrication, and artificial intelligence have propelled these devices beyond simple step counting toward sophisticated diagnostic and therapeutic platforms. This article reviews the latest research achievements in wearable sensor technology, highlighting key innovations in flexible electronics, multimodal sensing, and data-driven health prediction.

Materials innovation and flexible electronics

The fundamental challenge for wearable sensors has long been the mechanical mismatch between rigid silicon-based electronics and soft, curvilinear human tissue. Recent advances in stretchable conductors, self-healing polymers, and ultrathin substrates have largely overcome this limitation. Researchers at Stanford University demonstrated a fully stretchable sensor array using liquid-metal-filled elastomer channels that can conform to skin without signal degradation during bending or stretching (Kim et al., 2024). Meanwhile, the integration of MXene—a class of two-dimensional transition metal carbides—has significantly enhanced the sensitivity of strain and pressure sensors. A study published inNature Electronicsreported that MXene-based epidermal sensors achieved a gauge factor exceeding 500, enabling detection of subtle physiological movements such as pulse waveforms and vocal cord vibrations (Zhang et al., 2024).

Another breakthrough involves self-powered wearable sensors. Triboelectric nanogenerators (TENGs) now harvest mechanical energy from body movements to power continuous monitoring without external batteries. Wang and colleagues developed a textile-based TENG that converts walking motion into electrical signals while simultaneously measuring heart rate and respiration, achieving an energy conversion efficiency of 23% (Wang et al., 2025). This integration of energy harvesting and sensing represents a critical step toward autonomous wearable systems.

Multimodal and biochemical sensing

While physical sensors for heart rate, temperature, and motion are well-established, the recent frontier is real-time biochemical monitoring. Sweat, interstitial fluid, and tears have emerged as rich sources of biomarkers for glucose, lactate, cortisol, and electrolytes. A landmark study from the University of California, Berkeley introduced a wearable microfluidic patch that continuously analyzes sweat for multiple metabolites using enzymatic amperometric sensors (Gao et al., 2024). The device demonstrated high correlation with blood glucose levels in diabetic patients during exercise, opening avenues for non-invasive diabetes management.

Beyond sweat, microneedle-based sensors have gained traction for interstitial fluid analysis. A team at the Korea Advanced Institute of Science and Technology (KAIST) developed a graphene-coated microneedle patch that monitors both glucose and pH with a response time under 10 seconds (Lee et al., 2025). The microneedles are painless and dissolve after use, eliminating biohazard risks. Furthermore, researchers have expanded the biochemical repertoire to include stress hormones. A wearable cortisol sensor using molecularly imprinted polymers (MIPs) on flexible electrodes achieved detection limits as low as 1 nM, enabling real-time stress monitoring in clinical settings (Park et al., 2024).

Artificial intelligence and predictive analytics

The true power of wearable sensors lies not merely in data collection but in intelligent interpretation. Machine learning algorithms now transform raw sensor streams into actionable clinical insights. Deep learning models trained on multimodal wearable data—including electrocardiography (ECG), photoplethysmography (PPG), and accelerometry—can predict arrhythmias, hypertensive episodes, and even early signs of infection before symptoms manifest. A recent trial involving 5,000 participants showed that a convolutional neural network analyzing single-lead ECG from a smartwatch detected atrial fibrillation with 97% sensitivity (Attia et al., 2025).

Edge computing has further enhanced real-time analytics by processing data locally on the wearable device, reducing latency and preserving privacy. For instance, a custom-designed neural network chip embedded in a smart band can classify gait patterns and detect fall risk in elderly users within 50 milliseconds (Chen et al., 2024). Federated learning frameworks now allow multiple devices to collaboratively improve predictive models without sharing raw patient data, addressing critical privacy concerns.

Future outlook and challenges

Despite remarkable progress, several hurdles remain. Long-term stability of flexible sensors under repeated deformation and exposure to sweat continues to degrade performance. Biofouling—the accumulation of proteins and cells on sensor surfaces—reduces accuracy over days to weeks. Researchers are exploring antifouling coatings such as zwitterionic polymers and diamond-like carbon films to extend operational lifetimes.

Another challenge is standardization. The lack of universal calibration protocols for wearable biochemical sensors hinders their clinical adoption. Regulatory bodies, including the FDA, are working toward guidelines for continuous monitoring devices, but harmonization across different manufacturers remains incomplete.

Looking forward, the next generation of wearable sensors will likely integrate closed-loop therapeutic capabilities—sensing a biomarker and delivering a drug or electrical stimulation in real time. Early prototypes of "smart insulin patches" that monitor glucose and release insulin via microneedles have shown promise in animal models. Additionally, the convergence of wearable sensors with digital twins—virtual replicas of an individual's physiology—could enable personalized health simulations and preventive interventions.

In summary, wearable sensors have transcended their origins as consumer gadgets to become powerful tools for precision medicine. Through innovations in flexible materials, multimodal biochemical detection, and AI-driven analytics, these devices are poised to transform healthcare from reactive treatment to proactive, continuous wellness management. The coming decade will likely witness the seamless integration of wearable sensors into clinical workflows, ultimately empowering individuals with real-time insight into their own health.

References

Attia, Z. I., et al. (2025). Deep learning for atrial fibrillation detection using consumer-grade wearables.The Lancet Digital Health, 7(2), e112-e120.

Chen, L., et al. (2024). Edge AI for real-time fall detection using wearable sensor data.IEEE Transactions on Biomedical Engineering, 71(3), 789-798.

Gao, W., et al. (2024). A wearable microfluidic patch for continuous sweat metabolite monitoring.Nature Biotechnology, 42(5), 612-620.

Kim, J., et al. (2024). Stretchable liquid-metal sensors for conformal skin interfaces.Science Advances, 10(8), eadk4567.

Lee, H., et al. (2025). Dissolvable graphene microneedle patches for interstitial fluid glucose monitoring.ACS Nano, 19(1), 234-242.

Park, Y., et al. (2024). Molecularly imprinted polymer-based wearable cortisol sensor.Biosensors and Bioelectronics, 245, 115-123.

Wang, Z. L., et al. (2025). Textile triboelectric nanogenerators for self-powered health monitoring.Advanced Materials, 37(4), 230-241.

Zhang, Y., et al. (2024). MXene-based epidermal strain sensors with ultrahigh sensitivity.Nature Electronics, 7(3), 185-193.