Advances and Classification of Autonomous Systems in Biomedical Devices: Integration of Energy Harvesting and Ultra-Low Power Consumption

Abstract

Autonomous electronic systems are becoming increasingly important in people’s lives, as a result of advances in efficient energy storage systems, devices that can be permanently implanted in humans, and the trend towards compact devices that function as an extension of the human body. In addition, strategies continue to be found for the integration of energy harvesting in a constant and efficient manner. Covering the numerous advances made in biomedical devices can be quite overwhelming. This work presents a review of the latest strategies developed to produce energy from human body activity, the energy collectors for devices, and the strategies to create ultra-low-power wearable devices. The review focuses on the potential energy sufficiency required to power devices that can be implanted or worn, while also providing information about the patient’s condition. A comparison between the energies produced by different instruments and the improvements in the development of low-consumption devices is presented, with a focus on the type of medical devices. A new approach is established to classify and locate the most recent advances in autonomous systems in biomedicine based on their complexity/function.

1. Introduction

A medical device, as described by the FDA (the Food and Drug Administration in the United States of America), can range from a simple material to a complex electronic device, and may serve purposes such as diagnosing diseases or conditions, or curing, mitigating, treating, or preventing diseases. Medical devices are also defined as those intended to affect the structure or any function of the human body [1]. While many instruments fall within this definition, the intended focus of this review is on electronic devices that require energy and have the potential to function as autonomous systems. In [2], it is indicated that research is first focusing on producing low-energy-consumption devices, producing new alternatives for conventional batteries, and, finally, generating new ways to obtain energy. Currently, the development of biomedical devices is directed towards providing energy autonomy, seeking benefits such as portability, implantability, and energy savings. In this way, devices are being developed that do not have to be constantly recharged [3]. It has been expressed that, currently, a medical device must meet certain characteristics: it must be safe [3]; it must be minimally invasive [4], usable, or implantable [5]; and it must be autonomous and intelligent [5].

Methods have been developed to add energy autonomy to biomedical devices, including the use of triboelectric, thermoelectric, pyroelectric, and piezoelectric materials. However, the use of these materials does not generate enough energy to constantly power electronic devices. To increase energy production, strategies have been devised to optimize how these materials are used. The principle is to generate energy on a small scale for immediate use or storage. In recent decades, another type of Human Energy Harvester (HEH) device has gained relevance, which may be part of a medical device. Just as energy harvesting devices have been developed, new energy storage devices have also been developed, such as supercapacitors, which appear as an alternative for energy storage, since the use of batteries with constant and limited charge and discharge cycles causes the devices to lose the characteristics that interest us: portability and implantability. Furthermore, it is not only about developing self-sufficient autonomous systems, but also about developing ultra-low-consumption devices, to take advantage of new forms of energy generation. Analyzing the scientific and technological progress of these biomedical devices, several key areas are advancing separately but supporting each other, such as the development of techniques for generating and/or storing energy and designing low-energy-consumption instruments, with the objective of developing tools to make devices safe and capable of providing long-term treatment for human conditions. These devices have been classified in a wide variety of ways, such as a classification based on the type of risk, that is, on the potential impact on the patient’s health. They have also been classified based on the purpose, whether for diagnosis or treatment, where the diagnosis helps to identify a medical condition while the treatment is used to manage or cure some condition. Another classification is based on the application site, for example, the skin, the cardiovascular system, etc., or on the duration (short- or long-term), depending on the form of energy generation, such as thermoelectric, and triboelectric devices, among others.

In this review, a classification is made based on the complexity and function of the device. Therefore, the aim is to determine the progress of the different types of systems considered as biomedical devices that can be autonomous and intelligent and their areas of application. Moreover, the review seeks to establish their correlation with research focused on developing ultra-low-consumption devices. Special attention will be paid to those devices that are energy generators and how they can be classified. In addition, the correlation that exists between different technologies to extract energy from human beings will be investigated, for example, temperature, movement, and electromagnetic waves. Although it is true that there are numerous reviews addressing various aspects of these technological advancements, a new approach is established to classify and locate the most recent advances in autonomous systems in biomedicine based on their complexity/function. This approach enables a broader view of the current landscape and helps identify the direction of technological progress.

2. Nature of Generated Energy

In this constant advance of science, it has been found that some materials allow us to transform one type of energy into another; thus, transducers were born, capable of converting one type of energy into electrical energy, which is measurable and quantifiable, energy that can serve as a sensor for a process or as a generator of electrical energy to be stored or used in another process. Usually, the energy obtained in this type of system is through vibrations (triboelectric and piezoelectric), thermal gradients (thermoelectric generators), light fields (solar cells), and electromagnetic fields (rectennas), among others.

Figure 1 shows a general distribution of the main energy sources that are being investigated. This paper evaluates the forms of energy generation and how these are extracted from the environment under normal conditions within the daily life of a human being, and, also, which ones are intended to be applied to a usable device. These energies have in common that the electrical energy produced is not constant; it depends on the nature of how the energy is handled, meaning it can be stored or a signal conditioning circuit can be used to either rectify or regulate it.

3. Energy Harvesting Strategies

As mentioned in [6], there is no way to say which is the best method to collect energy, or how many types of energy could be acquired; it depends largely on the application of the system, the physical specifications, dimensions, and robustness, among others. When analyzing wearable sensors as devices to monitor a process or extract energy for humans, the sources of energy generation are limited. There are reports that energy is extracted from the temperature difference between the body and the environment [7,8]. In particular, research is seeking to develop more efficient methods for generating energy from movement, although this type of energy is limited, as typical human movements have frequencies below 5 Hz. Likewise, as a method to increase the energy produced, solar cells have been incorporated to extract solar energy, and, currently, the use of rectennas has been incorporated to obtain electrical energy from the electromagnetic energy present in the environment. When evaluating power production, these devices do not produce power consistently due to the nature of the signals being acquired and transformed into electrical energy, so it is necessary to incorporate a conditioning circuit to regulate and maintain the voltage. Kou et al. describe how wearable devices should address sustainability and stability needs and challenges while reducing the reliance on battery use [9]. Kou et al. also establish that a single source of energy capture can be inefficient, so strategies have been developed to group several energy sources, including sunlight, biomechanical energy, and temperature. In addition, new scenarios have emerged, such as the increase in the density of the world’s population, which has increased the use of mobile devices, giving rise to concepts such as the Internet of Things. The inclusion of cloud-based technologies implies the handling of large amounts of data wirelessly, which requires constantly increasing the quality of the data, which leads to working with more powerful technologies in radio frequency data transfer, a situation that has come to raise the opportunity to obtain energy from radio frequency signals found in the environment using rectennas as established by Kou et al. [9].

3.1. Strategies for Harvesting Energy from Movement

Strategies have been designed based on the type of energy collected. One of the most studied ones is the biomechanical movement; since movement is present in any environment that humans frequent, it is a promising way to capture energy through this physical phenomenon. Although triboelectric and piezoelectric sensors are well-known for their applications in vibration monitoring systems, these technologies can take advantage of the motion cycle for power generation. In particular, piezoelectric sensors are efficient at frequencies ranging from 60 to 100 Hz, unlike triboelectric sensors, which are efficient at lower frequencies [10]. Ways have been devised to extract energy through the oscillating movements of the human body, using electromagnetic energy and the energy generated by weight, as well as the friction caused by the movement itself. For example, Naval et al. have evaluated all possible uses of a triboelectric device to convert biomechanical energy produced by the human body into electrical energy, establishing a synergistic coupling so that the device extracts energy from each movement that the human body can perform [11]. Naval et al. report the results of simulations and experiments involving a device using a trinoboelectric generator (TEG) in four possible ways, after coupling the different ways these materials can be used to produce power and validate human movements, including translation, rotation, or torsion. Among the four power extraction modes evaluated, contact separation produced the highest maximum output power of 16.6 μW. The different operating modes of this device have different maximum power peaks, thus requiring different load resistors depending on their configurations and power production capabilities [11].

Hao et al. show that the different energy powers that can be generated with different parts of the body are established, and they also propose using the energy produced by the movement of the knee with a designed device [12]. In a similar way, another study shows that using a triboelectric nanogenerator (TENG) and a piezoelectric generator (PEG) can yield an output power of 4.1 W from the knee [13]. The device presented by Gao et al. is proposed to extract energy from knee movement [12], where the movement of a magnet is facilitated through a coil, and this movement is enhanced with a set of gears and a rope attached to one end to generate electrical energy, taking advantage of the natural movement of the body during movement. In this work, the device itself is used not only to generate electrical energy for devices but also as a sensor to detect the knee’s movement. In addition, in the work of Gao et al., the human-friendliness, adaptability, reliability, and economy (HARE) principle is cited in order to design a human energy harvester (HEH) device. The device was tested in walking and jogging people with a maximum output power of 66.85 mW, while the output power density reaches 0.07 W/kg. Pan et al. present a device such as a bracelet or anklet that is used to take advantage of the movement from the upper and lower limbs [14], and, also, a modeling of the structures is generated to improve the conversion of mechanical to electrical energy based on the frequency of the movement; in the same work, it is established that electromagnetic conversion has received special attention for its high power density characteristics, and, also, it is emphasized that human movement is highly irregular, so a device that converts this irregular movement into a regular one is necessary. Pan et al. discuss several strategies that have been used to capture the mechanical energy of the human body, which include the impact of the foot, the rotations of the joints, the energy of walking, the movement of the center of gravity, and the swing of the limbs. Finally, the article of Pan et al. deals with limb swing and cites several ways in which researchers have tried to capture this motion through various strategies.

3.2. Strategies for Harvesting Energy from Temperature

Thermoelectric generators are generally used in a planar manner [15], which limits their operation. Taking advantage of the deformation characteristics of their materials, a deformable configuration is established and tested to increase the efficiency of the material in energy production. Reference is made to other research indicating that, while increasing the legs of the thermogenerator increases its power, at the same time, it makes the material less flexible and, therefore, less usable in portable devices. In addition, the article presents the different topologies that have been developed to increase the efficiency of these materials. Furthermore, the work of Xing et al. presents the specific topology design methodology for deformable and flexible double-layer thermoelectric devices based on the Kriging-based material-field series expansion (KG-MFSE) method.

According to Erturun et al. [7], the incorporation of sensors in respiratory masks has increased in recent years. Among the parameters measured by these sensors are the heart and respiratory rate, oxygen saturation, ventilatory flow, environmental pollution levels, and VO₂max, among others. After describing the temperature difference between the inside and outside of the mask, a source of energy can be generated. In the work of Erturun et al., the use of an energy collector from a thermoelectric material is established, based on current needs, such as the KN95 mask. This personal protection equipment is known to be used almost permanently during the pandemic outbreak that occurred in 2020. From the temperature difference caused by human breathing, inside and outside the mask, a thermoelectric material, and a heat sink, Erturun et al. obtained temperature differences of 7 degrees Celsius. This resulted in an output power of ≈100 μW, with a corresponding power density of ≈30 μW/cm³. This, while the hot-side temperature was maintained at 31 °C and the output voltage of 70 mV, was recorded for cold-side temperatures of 24 °C.

Markiewicz et al. established the use of a thermoelectric device [8], where adds a special configuration to increase heat production based on vibrations, obtaining an improvement of 49%.

3.3. Hybrid Energy Harvesting Strategies: Integration and Technological Synergy

Shi et al. establish how research tries to improve energy production by joining two types of technologies [2]. They present a device to produce energy; although complex, it has a promising improvement in energy production. It is mentioned by Markiewicz et al. [8] that wearable devices are not only for the wrist as smart watches; now, they are in ears, chest, arms, and legs, among others. A good opportunity to implement these devices is in the human body, where they could take advantage of the constant movement and natural heat production. As mentioned in [8], their device generally establishes a fluid to cool one side of the thermocouple. Markiewicz et al. cite the work of Haras et al. [16], who demonstrate that moving the TEG between the heat source and the heat sink makes it possible to reduce the thermalization effect of the entire TEG, thereby increasing power and energy generation. This article proposes introducing a heat transfer medium between the heat source and the heat sink to prevent the movement of the TEG.

Shi et al. developed a device that combines two energy sources to generate enough energy for a wearable device [2], with a piezoelectric and an electromagnetic component, called a piezoelectric–electromagnetic energy collector. The device has a pair of piezoelectric elements and a pair of coils. The work of Shi et al. establishes how piezoelectric generators have been used and how designs have advanced to improve the efficiency of the devices in power generation, arriving at the concept of devices that increase the frequency of movement to increase efficiency and conversion-up. It also establishes the basic principle for energy conversion as electromagnetic energy from coils and permanent magnets together with a low-frequency movement can generate energy using the Faraday principle. They cite that their device can improve up to 23 times more energy than conventional collectors and improve 27% the overall power generation.

In [10], an analysis of the efficiency of PGs versus TENGs is carried out. One of the main focuses of attention in these autonomous systems is to increase energy conversion efficiency, where solar energy has an efficiency of 44%, and the typical efficiency of thermoelectric generators is 8–10%.

In [9], Kou et al. built a device that combines two energy harvesting technologies, triboelectric materials and the radio frequency collector (RF-TENG). Kou et al. establish the use of a hybrid device incorporating the use of a rectenna with a TENG material, together with a power management device, where this last part is the most difficult to include since the device has two energy sources, that of radio frequency waves that are of GHz frequencies, and that of TENG that obtains energy from movements with frequencies lower than 5 Hz.

Some works generate a design to collect energy from both a piezoelectric and a triboelectric sensor; the configuration of the materials and their mechanical design allow energy extraction from both forms simultaneously. In [17,18], they analyze the performance of one of these devices and compare the energy delivered by the piezoelectric model, by the triboelectric model, and by the hybrid model, where they highlight the capacity of the hybrid to generate energy.

4. General Model of an Autonomous Sensor

The basic principle in this type of devices is to enable systems to have the ability to work “perpetually”. Sometimes, a single device consists of several power generation sources, which intend to maintain a constant flow of current. In this case, an autonomous sensor is considered intelligent only when the acquisition and processing system is integrated within the physical design of the sensor [19].

Sometimes, they are accompanied by other electronic blocks, which depend on the application and function of the autonomous sensor. Sometimes, they include a circuit capable of transmitting and/or processing the detected information when necessary. Figure 2 shows the general structure and the different blocks that can be used to accompany this type of device.

5. From Simple to Complex Autonomous Systems

There is a wide range of simple and complex devices that can be considered as autonomous biomedical systems, generally consisting of a sensor, and they can include an analog–digital conversion, a data transfer, and processing and actuation, as well as a power supply that feeds all the parts that require it.

This type of work is found where a TENG generates energy from the movement of a rat’s diaphragm [20], and the generator is connected to electrodes, with the aim to energize the surrounding tissue and redirect the cellular growth of cardiac cells. In the same way, there are more complex devices, such as the one developed by Ma et al. [21]. That sensor was designed to be implanted inside a human without the need for a battery replacement, and with the ability to send various vital signs of the patient, while reporting the patient’s status wirelessly.

One of the objectives of autonomous biomedical systems is to increase the lifespan of devices by ensuring their energy source. An autonomous system must provide energy for itself; this means that it must be able to transform from one type of energy to another. This leads to defining an autonomous system as a device that can generate energy for itself, and, at the same time, it is capable of maintaining its operation in perpetuity for a specific purpose.

Figure 3 shows the different systems that, according to our references, can be considered autonomous, classifying them into four levels according to their complexity:

• Level 1: A device capable of generating its own energy and using it immediately to fulfill a specific purpose;
• Level 2: The integration of a sensor with an energy collector for storage or the sensor with a voltage conditioning system and its respective collector;
• Level 3: A device that generates power for itself and, at the same time, provides process information to a data acquisition and transmission system for analyzing and processing the resulting data on another device;
• Level 4: A system that is capable of producing and stably maintaining enough energy to process the information collected on site, which is already considered an intelligent system.

The systems considered in this work aim to generate energy constantly without the need to produce it from an external source. Although many investigations focus on the design of the collectors evaluating the multiple implementation options, with their respective energy analyses, the latest progress in the area of autonomous biomedical systems and the developments that are making this technological trend possible are detailed here. The electrical parameters of sensors reported in this study are presented in Table 1.

The system presented by Zhao et al. is a clear example of an autonomous sensor (self-powered sensing) [22], where the objective is to analyze the energy characteristics and the capabilities of the materials to function as sensors that do not require power and that can also provide more information than classic sensors.

5.1. Devices Level 1

Nanogenerators increase and direct cell development. Zhao et al. developed a device to help the maturation of cardiac cells [20]. Electrical stimulation of the tissue improves the distribution of cardiac cell stimulation produced by the nanogenerators.

An implantable device was developed by Li et al. [23] that generates direct current. It was implanted in the diaphragm of rats and supplied direct current through the movement of the diaphragm, which can function as a micro-power supply. This type of device meets the requirement of using batteries and allows us to have autonomous systems (implantable medical devices or IMDs). With reference to this work, pacemakers have an average consumption (power consumption) between 5–10 µW. Likewise, the material used is required to be biomechanically compatible with living tissues, using the measure of Young’s modulus, which was compared with its material, resulting in compatibility.

A piezoelectric nanogenerator is studied to measure blood pressure [24]. However, the authors do not introduce an energy storage system; an actuator turns on when the sensor registers high pressures; such a device could work as an alarm system. Cheng et al. indicate that this device could work with extra components as a wireless transmission medium for cellular devices. Like this example, many other studies are still in development, and, due to their current characteristics, they are classified as level 1.

A self-powered pressure-sensing system based on a conductive elastomer and flexible tf-TEG was developed for wearable applications [25]. According to Wang et al., high-thermal-conductivity materials in the form of a composite film and hydrogel heat sink are integrated to collect sufficient body heat to power the pressure sensor. This application using flexible tf-TEG sensors enables the excellent durability of more than 3000 cycles with a high response (24.9 ms) and high sensitivity regardless of the temperature gradients and internal resistance ratio, which find applications requiring real-time signal monitoring.

5.2. Devices Level 2

Dagdeviren at el. report a piezoelectric device that allows obtaining a signal conditioning stage [26]. This device generates energy from the movements of the rib cage, including the heart, lungs, and diaphragm. The device consists of several segments, including the piezoelectric sensor, a micro battery, and a rectifier.

5.3. Devices Level 3

5.3.1. Triboelectric

In the work of Zheng et al., an implantable triboelectric nanogenerator is presented [27]. The authors tested this device in in vivo conditions. The sensor has a multilayer structure, while it showed good performance and stability in vivo. When tested on an adult Yorkshire pig, electrical parameters achieved an open circuit voltage of 14 V and a short circuit current of 5 μA. This device has a means to store energy, a wireless transmission medium, and an autonomous sensor.

5.3.2. Piezoelectric

Zhang et al. developed a pacemaker with piezoelectric nanogenerator (PENG) technology [28], with characteristics such as biocompatibility, self-powered, good elasticity, and impermeability; and cardiac energy collection from the heartbeat. The authors evaluated the peacemaker on a canine model with elements such as the PENG, a voltage rectifier, a capacitor, reed switch, and wireless trigger.

5.3.3. Solar Energy

A prototype sensor with subcutaneous solar energy harvesting is proposed in [29], which is based on a wireless implantable device. The authors placed a solar panel under porcine skin to evaluate the sensor and exposed it to natural and artificial light sources. Depending on the light source, the panel generated power within the range of microwatts to milliwatts. The system includes a temperature sensor integrated into a node with wireless communication, a power management system, and energy harvesting capability.

5.3.4. Textile Sensor (TS)

Meng et al. report a wireless biomonitoring system based on a textile sensor (TS) for self-powered personalized healthcare with a high sensitivity of 3.88 V/kPa [30]. The wireless biomonitoring system could continuously monitor pulse waves throughout the night, even with body motion during sleep. In addition, this system could successfully diagnose obstructive sleep apnea (OSA) with greatly improved comfort and accuracy. The working principle of the TS for human pulse wave extraction encompasses two main aspects: the pulsating motion induces fiber deformation, and the deformation causes the generation of electrical signals. The TS can be directly sewn into different clothing places to measure pulse waves, for example, on the forehead, wrist, and chest, with aesthetic designs and comfort of use.

5.3.5. Temperature

Chen et al. present a device that can function as a solar energy collector while feeding a set of sensors for vital signs, temperature, heart rate, and oxygen saturation [31]. The data are transmitted via wireless Bluetooth to an IOT (Internet of Things) platform to monitor workers in environments that can facilitate heat stroke in hot weather. The device integrates several technologies to increase the efficiency of electric generators, keep the device running, and transmit data during its active operation time.

5.4. Devices Level 4

For noninvasive real-time human heart rate monitoring, Lin et al. developed a self-powered wireless body sensor network (BSN) system for heart rate monitoring after integrating a bottom-frame-based triboelectric nanogenerator (D-TENG), a power management circuit, a heart rate sensor, and a signal processing unit [32]. The heart rate signal acquired by the sensor was processed in the signal processing circuit, sent to an external device via the Bluetooth module, and displayed on a personal mobile phone in real time. This work presents a competent and cost-effective solution to meet the increasing demand for daily health monitoring. By converting the inertial energy of human walking into electrical energy, the generator delivered a peak power of 2.28 mW with a total conversion efficiency of 57.9% at a low operating frequency, which is capable of immediately and sustainably driving the highly integrated BSN system. While the number of reviews on these materials is growing, the development of research of this type is also increasing, as mentioned in [33]. Such growth is a consequence of the increase in the use of triboelectric materials, which has been growing since 2010. It reinforces the idea that there is an interest in this type of material and the applicability that the generated energy could have.

Table 1. Table showing research studies illustrating their currents (Isc), voltages (Voc), and power (power and power density). ND: no data.

Device	Sensor Site	Voc	Isc	Power	Power Density 1	Mechanism	Ref.
Implantable	Inside the thoracic cavity	8 V	0.8 μA	ND	20 mW/m2	Tf-TEG	[23]
Wearable	ND	3.3 V	0.6 μA	ND	10.4 mW/cm3	Flexible TENG	[34]
Implantable	Inside the thoracic cavity	75 V	12 μA	ND	ND	Flexible TENG	[21]
Implantable	Inside the body	540 V	42 µA	2.3 μW	7.42 Wm2	ND	[22]
Implantable	iTENG in vitro, in vivo	78 mV	ND	7.9 µW	ND	Tf-TENG	[25]
Implantable	In vitro, exvivo	4.04 V	6.16 × 10−8 A	ND	ND	Mecánica PN	[33]
Implantable	Ex vivo	14 V	5 µA	ND	107 mW/m2	i-TENG	[27]
Wearable	On the skin	430 mV	75 mA	ND	ND	Solar-TENG	[31]
Wearable	Over clothing	79 mV	5.6 to 12 μA	2.28 mW	ND	D-TENG	[32]

¹ peak output power density, electrical output achieved a peak voltage.

Table 1. Table showing research studies illustrating their currents (Isc), voltages (Voc), and power (power and power density). ND: no data.

Device	Sensor Site	Voc	Isc	Power	Power Density 1	Mechanism	Ref.
Implantable	Inside the thoracic cavity	8 V	0.8 μA	ND	20 mW/m2	Tf-TEG	[23]
Wearable	ND	3.3 V	0.6 μA	ND	10.4 mW/cm3	Flexible TENG	[34]
Implantable	Inside the thoracic cavity	75 V	12 μA	ND	ND	Flexible TENG	[21]
Implantable	Inside the body	540 V	42 µA	2.3 μW	7.42 Wm2	ND	[22]
Implantable	iTENG in vitro, in vivo	78 mV	ND	7.9 µW	ND	Tf-TENG	[25]
Implantable	In vitro, exvivo	4.04 V	6.16 × 10−8 A	ND	ND	Mecánica PN	[33]
Implantable	Ex vivo	14 V	5 µA	ND	107 mW/m2	i-TENG	[27]
Wearable	On the skin	430 mV	75 mA	ND	ND	Solar-TENG	[31]
Wearable	Over clothing	79 mV	5.6 to 12 μA	2.28 mW	ND	D-TENG	[32]

¹ peak output power density, electrical output achieved a peak voltage.

6. Ultra-Low-Power Devices Developed

The development of ultra-low-power devices is crucial for the advancement of autonomous sensors; as seen up to this point, the energies generated by these sensors can be constant but are not of great amplitude, so the integration of miniaturization technologies is being promoted [35], as well as the development of increasingly smaller integrated circuits and ultra-low-power wireless communications. The development of autonomous systems has driven strategies to generate ultra-low-power circuits, improved energy harvesting strategies, and generated an improvement in wireless data transfer through ultra-low-power communications, where all these technological advances have allowed the development of smart sensors.

Jha et al. present a very-low-power biomedical frontend designed in 0.18 μm CMOS process technology [36]. It can be used as a self-powered acquisition system, and, with the use of the wireless body area network (WBAN), this system is suitable for ubiquitous healthcare (u-Health). The system has a signal conditioning stage with programmable gain and bandwidth, a mixed-signal automatic gain control (AGC), and a ΣΔ ADC. The ADC provides ~12 bits and the overall system consumes ~2 μW of power. It was originally conceived for the continuous monitoring of ECG biosignals.

7. Energy Storage Systems

As detailed by Azega et al. [37], there is a strong trend in research aimed at improving the characteristics of batteries and supercapacitors as energy storage devices; currently, part of the electric automotive industry depends mainly on these characteristics.

In this regard, the work of Azega et al. highlights technological advances related to supercapacitors and rechargeable batteries [37]. On the one hand, that report shows the characteristics of both devices and, in particular, their differences; on the other hand, they establish the areas of interest of the research between the two currents, where supercapacitors science is focusing on discovering new ways to maximize their efficiency, on the modification of the electrode materials, and creating new designs. In the same way, Azega et al. present the areas of interest in the development of batteries, with a particular focus on the selection of the cathode and the anode. These parameters will improve the performance of batteries and generate safer devices since lithium-ion technology, under certain circumstances, can become unstable.

The literature shows us that lithium-ion batteries are excellent at generating large energy discharges; they have a high energy density. However, maintaining these discharges is not their strong point; supercapacitors have a high power density, which allows them to retain power for long periods, in addition to the fact that the wear of these devices is less than that generated by batteries [38,39,40].

8. Discussion

This article provides an overview of the advances in the development of autonomous biomedical devices found in the literature, emphasizes the trends directing the generation and design of these devices, and proposes a classification based on the complexity of the autonomous systems reviewed.

As can be seen in general terms, energy harvesters today can autonomously generate energy for devices that require instantaneous power at the microwatt to milliwatt levels, depending on the technology and harvesting strategy used, without neglecting the importance of the environment in which the device is used. Devices that generate higher levels of current and power integrate several forms of energy generation; it is also observed that devices with a better-developed strategy to extract energy improve the amount of energy produced. A recent review describes how piezoelectric and triboelectric materials are used to develop implantable biomedical devices [41]. Such results highlight a current research trend in using these materials as energy generators that can be implanted in living organisms.

As mentioned in the previous sections, some designs of biomedical devices considered autonomous systems do not require batteries. The energy source can provide a constant current; however, in the vast majority of collectors developed, the energy source is not continuous, so they require a means to store energy when this source is needed. This situation makes integrating energy storage devices, such as supercapacitors and batteries, necessary.

From the work of Daoud et al. [42], we see trends supporting the integration of energy collectors and energy regulation systems with low-consumption devices. In their work, the authors present the design of an intelligent system to manage energy storage sources using batteries and supercapacitors to improve the battery life. The biosensor is counting on a sensor that consumes 1.5 mW and a wireless data transfer system.

In addition to focusing on efficiency, the integration of supercapacitor designs into ultra-low-consumption devices and energy harvesters is being explored. In [43], Pantrangi et al. describe how flexible micro-supercapacitors are designed for wearable devices as a result of their high power density, superior power density, flexibility, and safety characteristics. Undoubtedly, research is heading towards discovering the technology needed to generate autonomous and intelligent biomedical systems that improve people’s quality of life.

Often, research efforts focused on developing implantable or wearable devices do not prioritize the costs of materials or the equipment needed for mass production. Typically, these aspects are addressed in earlier studies that test and analyze the energy generation capabilities of specific materials. However, we did find studies that consider the industrialization of these devices. For instance, Fan et al. said [34], “The entire fabrication process does not require expensive raw materials or sophisticated equipment, which would benefit mass industrialization.” Nevertheless, we believe such clarifications are crucial and should be explicitly stated in studies of this nature.

When discussing mass production, being informed about the regulatory requirements involved is essential. Most of the reviewed studies do not explicitly mention this necessity, though this does not mean it is overlooked. As researchers and developers of software or hardware, we often focus on identifying and addressing a need. At the same time, aspects such as sales, mass production, and commercialization are frequently secondary considerations. However, developing a product that solves a problem without taking into account the scenarios for which it is intended can lead research efforts to dead ends. Khan et al. mention that commercialization and mass production are complex topics and come at a high cost [41], particularly for implantable devices—a statement we fully agree with. One of the main challenges in these studies lies in the rapid advancement of technology and the lengthy processes required to bring innovations to market, which include technology development, patent authorization, and the publication of related technologies. Currently, with the use of artificial intelligence and machine learning models, science is advancing at a remarkable pace. However, many creations remain stalled due to regulatory procedures, preclinical and clinical trials, compliance with various standards, and certifications. We believe that commercialization policies are perhaps the most significant challenge. These aspects are not to say they are unnecessary, but rather that technological advancements are outpacing these processes.

9. Conclusions and Future Directions

This review presents a broad overview of the different advances that we can find in the literature regarding the development of autonomous biomedical devices. This paper presents a new way of classifying devices. In addition, this report shows how some research will gradually mature over time. Autonomous medical devices are heading toward the development of self-sustaining, implantable, and intelligent devices.

Likewise, energy generation depends on the surrounding environment where a human operates, and the energy collection strategies limit the quantity and amplitude of energy that can be collected. However, considering any of the four mentioned levels, low-consumption devices can obtain sufficient power levels.

The current research focuses on developing autonomous and intelligent biomedical devices that optimize energy use and provide users with a more comfortable and safe experience. With the development of energy harvesting, management, and storage technologies, we are getting closer to a new generation of devices combining efficiency, autonomy, and utility to improve people’s lives.