The fourth industrial revolution, also known as Industry 4.0, has revolutionized the healthcare industry by enabling the widespread adoption of health monitoring sensors (HMS). These sensors are characterized by their digitalization and intelligence, offering a wide range of applications in medical care, personal health management, elderly care, and sports. They provide people with more convenient and real-time health services, transforming the way we monitor and manage our well-being.
However, these sensors are not without their limitations. They can be subject to noise and drift, leading to fluctuations in their measurements. Additionally, the increasing availability of sensors has resulted in the collection of large amounts of data, making it challenging to extract meaningful insights and information. Traditional sensors are also often characterized as open-loop, lacking feedback or control signals from the systems they monitor, limiting their ability to directly impact the behavior of the systems they observe.
The development of artificial intelligence (AI) has emerged as a powerful solution to address these limitations. AI provides advanced tools and algorithms for data processing and analysis, enabling intelligent health monitoring and achieving high-precision predictions and decisions. By integrating the Internet of Things (IoT), AI, and health monitoring sensors, it becomes possible to realize a closed-loop system with the functions of real-time monitoring, data collection, online analysis, diagnosis, and treatment recommendations.
Advancements in Wearable Sensors
The continuous advancement of wearable electronics towards multifunctional wearable systems is driven by the desire to improve the quality of life through the enhancement of external device functionality. Commercially available wearable devices, such as wristbands, watches, and glasses, typically consist of rigid elements with flexible belts that are worn on the human body. However, to further optimize wearable comfort and enable advanced healthcare interactions with humans, wearable electronics are now progressing towards platforms with excellent flexibility, stretchability, and even self-healing capability, benefiting from significant advancements in the development of functional flexible materials.
Wearable sensors have made their way into the field of digital health, finding diverse applications in biomedical settings. These sensors enable the monitoring of vital signs, such as respiration rate, heart rate, blood pressure, skin temperature, and pulse, as well as physiological signals, such as electrocardiography (ECG), electromyography (EMG), and electroencephalography (EEG). They can also capture body kinetics, such as strain and pressure, as well as dynamic biomolecular states through accessible biofluids like sweat.
Innovative Wearable Sensor Designs
One notable example of an innovative wearable sensor is the Tunable Ultrasensitive Nature-inspired Epidermal Sensor (TUNES), developed by researchers. This sensor mimics the geometry and tuning mechanism of the slit organ in spiders, exhibiting the capability to detect a broad range of signals ranging from minute pulses to more substantial muscular contractions and respiration. By creating nanoscale cracks on a metalized polyimide film and adjusting the sensor’s sensitivity through pre-strain, the TUNES can measure strain and pressure across various scales using a single sensor.
Another example is a respiratory rate sensor fabricated on a face mask through oxidative chemical vapor deposition (oCVD). This sensor utilizes a highly conductive Poly(3,4-ethylenedioxythiophene) (PEDOT) layer to extract blood pressure information and respiratory rates in real-time with excellent precision. The sensor works by measuring the changes in current associated with the wearer’s respiration, detecting inhalation and exhalation cycles as current drops and recoveries, respectively, and measuring the duration and amount of current change for each cycle to estimate the respiratory rate.
Temperature sensing also plays a pivotal role in delivering critical data in various scientific and engineering domains. Researchers have developed a highly sensitive and flexible artificial skin with a negative temperature coefficient (NTC) thermistor that can be conformably attached to the facial surface, enabling continuous and accurate measurement of physiological temperature over long periods. This sensor can capture small temperature variations associated with inhalation and exhalation, as well as monitor changes in respiration patterns during intense physical activities.
Accurate recording of human biopotential signals, crucial for diagnosing and treating heart, brain, and muscle-related diseases, relies on the use of efficient wearable electrodes that interface effectively with the skin. Researchers have developed a highly conductive polymer dry electrode (PWS) film that possesses remarkable attributes, such as self-adhesiveness, stretchability, and conductivity, enabling the acquisition of high-quality signals for ECG, EMG, and EEG under diverse conditions.
Advancements in Chemical Sensing Wearables
Obtaining a comprehensive assessment of an individual’s health status requires collecting extensive information, including vital signs, physical activities, and chemical biomarkers from or near the human body. Wearable chemical sensors have been developed for the real-time detection of biomarkers from biofluids, such as tears or sweat, which could potentially enhance disease prediction, screening, diagnosis, and treatment.
One example is a wearable patch designed to continuously track sweat rate during rest periods. To address the challenge of low sweat secretion rates during rest, microfluidics are integrated into the patch, preventing evaporation and detecting sweat rate selectively. Additionally, a laminated hydrophilic filler is incorporated, enabling rapid uptake of sweat into the sensing channel, reducing sweat accumulation time.
Contact lenses have also garnered significant attention as a viable substrate for tear sensors due to their biocompatibility and compliance. Researchers have introduced a tear sensor using a graphene field-effect transistor (FET) integrated with a graphene-silver nanowire composite antenna, which works at radio frequency and wirelessly transmits sensory information upon the occurrence of glucose oxidation on the graphene channel.
Wireless Interconnection and Energy Management
Establishing a functional network for wearable devices poses challenges, such as disruptions in physical activity caused by direct wiring between sensor nodes. Near-field communication (NFC) offers an alternative approach where sensors can be wirelessly powered by a reader, enabling battery-free and secure operation for various physiological measurements on the skin or inside the body.
Researchers have also developed an innovative technique for constructing a wearable sensing system on textiles using metamaterial textiles to enable the propagation of radio surface plasmons on the body surface, leading to a secure and energy-efficient wireless body sensor network. This approach employs conductive fabrics in clothing that facilitate surface-plasmon-like modes at radio communication frequencies, resulting in body sensor networks with significantly higher transmission efficiencies than traditional radiative networks without metamaterial textiles.
Advancements in Implantable Sensors
Compared to wearable sensors that monitor body surface markers, epidermal electrical signals, and body movements, implantable sensors function inside the body and can perform more direct monitoring of healthy states. For example, they can monitor the normal functions of the heart and blood vessels, the recovery progress of injured tissue, and the abnormality of the central nervous system.
The research on implantable devices also makes high requirements for materials selection, structure design, system integration, power supply, and biological safety. In recent years, the development of self-powered implantable sensors has exhibited the potential to revolutionize healthcare by providing continuous and accurate data without the need for external power sources or frequent replacements.
Self-Powered Cardiac Monitoring
One example of a self-powered implantable sensor is the implantable active pressure sensor (iTEAS), which is based on the principle of the triboelectric effect. iTEAS can convert mechanical signals generated by motion from the implant site into readable electrical signals for further analyzing heart rates, blood pressure, blood flow, and respiratory rates. Simultaneously, iTEAS can harvest mechanical energy from the human body and convert it into electrical power to self-supply, overcoming the limitations of battery capacity.
Another device, the self-powered pressure sensor (SEPS), is a novel approach to monitoring intracardiac pressure using a minimally invasive implanted heart catheter. SEPS achieves a sensitivity of 1.195 mV/mmHg with a linearity of R^2=0.997, addressing the limitations of traditional invasive monitoring methods by providing a minimally invasive and self-powered solution.
Biodegradable and Transient Sensors
Researchers have also developed a bioresorbable triboelectric sensor (BTS) for use in cardiovascular postoperative care. Based on the triboelectric effect, BTS is designed to be implanted in the body and is capable of monitoring pressure changes in real-time. BTS has the potential to improve patient outcomes by providing accurate and continuous pressure monitoring without the risks associated with permanent implants.
In addition to cardiac monitoring, researchers have introduced a control system for monitoring bladder pressure and controlling urination. This system employs a TENG-based pressure sensor to detect the filling state of the bladder and a bistable micro-actuator based on shape memory alloy to induce contraction and relaxation of the bladder for urination. This approach provides a reliable solution for future clinical applications.
Neural Interfaces and Neurotransmitter Sensing
The nervous system is the human body’s control center, regulating organ activity and mediating responses to external stimuli. Monitoring neural electrical signals or neurotransmitters provides insights into physiological and pathological processes, informing the prevention and treatment of neurological disorders.
Researchers have developed a flexible and biocompatible neural ribbon electrode (NRE) that achieved self-adaptation to various diameter nerves by wrapping around nerve fibers. The NRE, coated with electrically-stable carbon nanotubes, ensures a close 3D non-invasive contact with neural tissue and stable communication, enabling the recording of neural signals from small nerves in a non-invasive way.
Another innovative design is the spiked ultra-flexible neural interface (SUNI), which collects sensory information from rat mechanoreceptors through its spike structure. The novel 3D structural design of SUNI enables well-conformal contact with nerve fiber bundles, contributing to high-quality recordings that differentiate tactile and proprioceptive stimuli, providing high spatial resolution classification of neural signals.
To address the mismatch between hard and hydrophobic neural interfaces and the soft and wet nervous tissue, researchers have developed various strategies, including designing intrinsically hard micro-silicon-based devices with a flexible link structure, using flexible materials as a bridge between implants and biological tissues, and developing intrinsically stretchable implants based on advanced soft materials.
AI-Enhanced Sensor Designs and Data Processing
The integration of AI into sensor technologies holds immense promise for the future of healthcare and biomedicine. AI-enhanced sensor systems can enable personalized healthcare by continuously monitoring physiological parameters and integrating them with an individual’s medical history to generate personalized recommendations for disease prevention, early intervention, and chronic disease management.
AI-Assisted Sensor Design
AI algorithms can facilitate the automated design of volatile organic compound (VOC) sensors, eliminating the need for arduous and time-consuming design processes. This is exemplified in the design of surface-enhanced infrared absorption (SEIRA) antennas, where the analysis of the analyte molecule’s infrared spectrum and the subsequent selection of an appropriate antenna structure to match molecular vibrational frequencies are critical steps. AI-assisted design systems can efficiently handle these tasks, leading to more effective and efficient design outcomes.
Researchers have also demonstrated the use of inverse design techniques, where semi-supervised deep learning algorithms are employed to encode the structural design and optical responses from input geometry, enabling the reconstruction of the original structural geometry and the generation of various output candidates for reverse design.
AI-Driven Data Processing
AI algorithms can also facilitate the analysis and processing of data from VOC sensors in a prompt, direct, and automated manner. Principal component analysis (PCA) and support vector machines (SVMs) are among the techniques used to reduce the dimensionality of spectral data while preserving essential features, leading to a decrease in the amount of data, simplification of data processing, and prompt generation of test results.
Researchers have demonstrated the integration of PCA and SVM algorithms to achieve 100% recognition accuracy for the identification of methanol, ethanol, and isopropanol using a hook-shaped nanoantenna array that utilizes wavelength multiplexing for continuous broadband detection of multiple absorption peaks in the fingerprint region.
Toward Comprehensive Body Sensor Networks and AIoT
The rapid development of sensor technology has been a driving force behind Industry 4.0. The integration of AI data analytics with wearable sensors allows for the capture of crucial information, such as muscle deformation, joint bending, temperature changes, and heartbeat frequency, which is invaluable for a wide range of applications, including healthcare, environmental monitoring, human-machine interactions, and smart cities.
Multimodal Sensor Fusion for Enhanced Functionality
Researchers have demonstrated the benefits of multimodal sensor fusion, where the integration of visual data from cameras and somatosensory data from skin-like stretchable strain sensors can achieve a stellar recognition accuracy of 100% for human gesture recognition, maintaining this accuracy even under less than ideal conditions for the image sensor, such as noise, under-exposure, or over-exposure.
Another innovative method of data fusion from multiple sensors uses a hierarchical SVM (HSVM) algorithm, which merges the outputs from pressure sensors and radar technology to achieve a classification accuracy of 92.5% for hand gesture recognition, outperforming standalone pressure sensors and radar technology.
Intelligent Wearable and Implantable Systems
Beyond wearable sensors, researchers have also explored the integration of tactile sensors, triboelectric sensors, and machine learning to create intelligent wearable devices for various applications, such as:
- Smart gloves with multimodal sensing and augmented haptic feedback for gesture recognition and control in virtual and augmented reality environments.
- Self-powered and intuitive glove-based human-machine interfaces that combine superhydrophobic triboelectric textile sensors with machine learning for complex gesture recognition and control.
- Sign language recognition and communication systems based on smart glove sensors and deep learning algorithms for seamless translation between sign language and speech.
In the realm of implantable systems, researchers have developed self-powered and AI-enabled floor monitoring systems that can accurately detect an individual’s position and identity through unique gait patterns captured by triboelectric sensor arrays. These systems address privacy concerns associated with camera-based monitoring, showcasing the potential of AIoT-driven solutions for healthcare, security, and automation applications.
Conclusion
The advancements in sensor technologies, particularly the integration of AI, have opened up new horizons in healthcare and biomedical applications. The fusion of flexible, wearable, and implantable sensors with intelligent data processing and wireless communication capabilities has led to the emergence of comprehensive body sensor networks and AIoT-driven solutions.
These innovations enable continuous and pervasive monitoring of an individual’s health status, both inside and outside the body, empowering patients to actively participate in their healthcare management and fostering a proactive approach to well-being. The convergence of AI, self-sustainable IoT systems, and advanced therapeutic devices holds great promise for the development of closed-loop sensing-therapy systems, providing enhanced benefits to patients suffering from chronic diseases.
As the field of AI-enhanced sensors continues to evolve, addressing challenges related to data privacy, security, regulatory compliance, and robust validation will be crucial for building trust and enabling the widespread adoption of these transformative technologies in healthcare and biomedicine. By overcoming these hurdles, the integration of AI and sensors will pave the way for personalized, predictive, and preventive healthcare, revolutionizing the way we approach medical diagnosis, treatment, and management.
Sensor Networks is at the forefront of these advancements, providing a platform for the latest research, insights, and innovations in the realm of sensor networks, IoT, and related technologies. Explore our comprehensive resources to stay informed and engaged with the rapidly evolving landscape of sensor-enabled solutions that are shaping the future of healthcare and beyond.