I. Introduction

Technological advancements have always been one of the important key drivers for innovation in healthcare. The evolution of devices for wearable and remote healthcare over the last 2 decades is just one clear example. Indeed, fueled by innovations in sensor technology, material technology and micro- and nanoelectronics, various types of medical sensing technology previously only possible in a hospital environment are now readily available in wireless wearable devices. Wearables are now available as medical devices, but it is equally interesting to see how a wide array of consumer products are now available that allow an individual to measure vital signs like heart rate, respiration rate, blood oxygenation, etc. in a semi-continuous manner.

To a large extent, the development of wearable technology is driven by medical challenges in the cardio-vascular space. Indeed, most wearable devices focus on measuring and interpreting data related to the autonomous nervous system, the cardiovascular system, the pulmonary system and to a lesser degree the central nervous system. Existing personal/wearables are mostly unable to measure the gastro-intestinal system. Unfortunately, globally metabolic health is dropping at an alarming rate. Metabolic disorders occur when the normal chemical reactions are disrupted, resulting in either too many or too little of critical substances. While some are genetically inherited, others are developed when critical organs, like liver, pancreas or bowel are diseased. But also, lifestyle, behavior and nutrition are critical for our metabolic health. For a lot of digestive disorders, the exact pathology and underlying disease mechanisms remain not fully understood. In an effort to address these challenges, various research groups around the world are betting on advanced ingestible technology that could allow unique and unprecedented insight into the human metabolic system^[1]. It is clear that one of the areas where we might see a similar explosive growth, is ingestible sensor, as shown in Fig. 1(a), focused on demystifying our gastrointestinal inner workings.

Fig. 1(b) shows the block diagram of the ingestible sensor. The radio, which includes a matching network, transmitter (TX), receiver (RX), and Local oscillator, enables communication with the outside world. The signal detected by the sensor is amplified, filtered, and converted into digital data through the sensor Analog Front End (AFE). This data is stored in memory and periodically transmitted to an external RX via TX. Control signals from the outside are received through the RX and used to control the Micro Controller Unit (MCU). The MCU then controls the ingestible sensor. The matching network ensures impedance matching between the radio and the antenna, and the local oscillator generates the carrier frequency. The Power Management Unit (PMU) manages the required power for each block. The Advanced High-performance Bus (AHB) enables data transfer between blocks.

The biggest challenges for the envisaged applications are the size constraints and the tissue depth. There will be innovations needed to push the device size down, while ensuing still reliable operation for a free-floating ingestible sensor. To do so, ultra-low-power wireless communication with a small footprint, battery and antenna size is required.

In this paper, wireless system miniaturization solutions for ingestible sensors are discussed. Some of the technical challenges in miniaturization are introduced. State-of-the-art miniaturized wireless systems are reviewed, and future trends are forecasted.

This paper is organized as follows. State of the art ingestible sensors and its wireless systems are introduced in Section II. Section III discusses miniaturization techniques for wireless systems. Section IV suggests the future trends in wireless systems in ingestible sensors. Finally, our conclusions are drawn in Section V.

Fig. 1. (a) Ingestible sensor that connects the inside of the GI tract to the outside world; (b) its block diagram.

II. Wireless Systems in Ingestible Sensors

Advances in integrated circuits, RF wireless communication, power management and sourcing lead to a quantum leap in the evolution of ingestible sensors, enabling miniaturization and low power consumption while wireless connectivity. Capsule-shaped, ingestible camera pills^[2], are clear examples. In addition to image sensing, ingestible sensors detecting biomarkers are being actively researched. These existing sensors are summarized in Table 1.

After the PillCam^[2]’s successful launch, numerous companies provide ingestible camera pills for endoscopic. These capsule endoscopies travel through the gastrointestinal tract, transmitting the imaging data to an external RX wirelessly. Such capsule shaped wireless optical sensors eliminate the need for sedation and invasive endoscopic insertion and enables the imaging of the small intestine, which is difficult to access with conventional endoscopic endoscopies.

The gastrointestinal (GI) monitoring with images has significant advantages, but it presents notable limitations. Firstly, high-resolution image sensing, requiring data transfer rates of tens of Mb/s is demanded for reliable accurate diagnostics^[11],^[12]. It is problematic to secure the communication range through the tissue from capsule to the base station. An additional device as a ``repeater'' attached on the skin, including external transceivers and antenna array, is required^[7],^[9],^[11],^[12].

Secondly, the limited scope of diseases detected by visual inspection and the inability to assess nutritional status make it unsuitable for a comprehensive diagnosis of gastrointestinal health. Third, the capsule endoscope has a short battery life within 12 hours^[2], due to the use of high data rates and LEDs, too short to diagnose gut motility, monitoring of digest, chronical diseases. Due to the limitations, the camera pill is hard to provide multi-modal sensing, required for the comprehensive assessment of the digestive system where various and complex physiological processes occur.

As an alternative, biomarker based ingestible sensors that monitor the GI tract via electrochemical sensors have been recently introduced^[13]. Unlike optical/image sensors, the biomarker sensors can mitigate the data rate (< 10 Mb/s), while lowering the power consumption (< 10 mW). Therefore, the biomarker based ingestible sensors can provide more reliable wireless communication, e. g., longer communication range, direct connection to off-the-shelf devices without the need for repeaters. This can also increase the communication link budget, relaxing the power budget to the wireless module or minimizing the antenna size to further optimize the whole system form factor.

Biomarkers such as body temperature, pressure, pH, ion, oxidation-related, biomolecular, gases are measurable indicators of a biological condition or disease condition. A biomarker sensing based ingestible sensor in^[13] detects the changes of pH, temperature, and pressure via sensors, when traveling through the GI tract, for measuring the transit times through the stomach, small intestine, colon, and the entire digestive. This provides a comprehensive profile of gut motility, aiding in the diagnosis of conditions such as gastroparesis and chronic idiopathic constipation.

Ion sensors utilizing Ion-Selective Electrodes (ISE)^[14],^[15] sense dietary minerals such as Na+, Ca2+, Mg2+, and K+ in GI fluid, enabling the evaluation of electrolyte balance in the GI tract, as illustrated in Fig. 2(a).

Oxidation-related biomarkers are used for diagnosing intestinal diseases such as Inflammatory Bowel Disease (IBD)^[16],^[17]. The biosensor bacteria^[16] or the chemiluminescent paper-based sensor^[17] realized the oxidation-related biomarkers. They express luminescence when exposed to certain inflammation-related markers, and this light is converted into electrical signal by photodetector, as shown Fig. 2(b).

Fig. 2(c) shows the biomolecular sensor that utilizes a 15-pixel CMOS fluorescence sensor array for biomolecular sensing^[18]. Each pixel in the sensor array is to react to specific target biomolecules such as DNA or proteins. The sensor array is capable of sensing various biomarkers simultaneously in GI fluid and thus collects diverse responses of the gut from new drugs or treatments, accelerating medical advancements.

Gas sensing has been also introduced to monitor the GI^[20], identifying gases such as oxygen (O$_{2}$), hydrogen (H$_{2}$), and carbon dioxide (CO$_{2}$) that serve as indicators of various biological and chemical processes in the gut, as depicted in Fig. 2(d). This sensing method allows to distinguish areas of the intestine and understand the fermentation patterns of intestinal microorganisms through the thermal conductivity and resistance changes on the sensor surface in response to different gases. Apart from sensors for detecting biomarkers, there are ingestible sensors that react to specific drugs to verify medication intake^[23].

Fig. 2. Biomarker based sensors: (a) Ion Sensing ^[15]; (b) Oxidation ^[17]; (c) Fluorescence Sensor Array ^[18]; (d) Gas sensing ^[20].

The required data transfer rate for ingestible sensors aforementioned is dependent on the type of sensors. This ranges typically from tens of b/s (temperature sensor, blood sensor, etc.)^[24] to several Mb/s for more complex sensing applications such as oxidation (1 M/s,^[17]), Gas (1 Mb/s,^[20],^[22]), Ion sensing (4 Mb/s,^[14]), and biomolecular sensor (7 Mb/s,^[18]). Furthermore, high-resolution image sensing requires even higher data rates. When the data rate increases, the data communication range decreases, according to Shannon’s theorem^[25], thus requiring more power to maintain the communication range. For communication of ingestible sensors, data rate, distance, power consumption must be considered depending on the sensor type, purpose, environment.

These requirements present other considerations such as selecting frequency band, standard, tissue loss, antenna's radiation pattern change, co-existence with other sensors, miniaturization.

Selecting the frequency band and standard protocols for communication of ingestible sensors is crucial. Commonly used Industrial, Scientific, and Medical (ISM) bands, e. g., 2.4 GHz, accommodate several standards such as Wi-Fi^[26], Bluetooth^[27], Bluetooth Low Energy (BLE)^[28], and Zigbee^[29]. Bluetooth and Wi-Fi are attractive standards for the ingestible sensor to be compatible with off-the-shelf IoT devices, but their relatively high power consumption (100 mW ~ 800 mW) would be problematic^[24],^[30]. BLE designed for intermittent communication, i. e., duty-cycling, lowers power consum-ption but has limitation to continuous data collection and real-time monitoring. Zigbee, reported as a lower power consuming standard than Bluetooth^[31], is limited for sensors that requires high data rate (~ Mb/s), as it supports only up to 250 kb/s.

The 2.4 GHz ISM band has the advantage of using diverse standard protocols that is highly compatible with a wide range of off-the-shelf mobile devices, but high susceptibility to interference and weakness to security due to its high congestion. In addition, the 2.4 GHz ISM band has relative high tissue loss compared to other lower frequency bands. For example, with the tissue thickness of 7.8 cm, the loss is about 90 dB at 2.4 GHz, whereas it is about 40 dB at 400 MHz^[32].

Therefore, the lower frequency band, the 402-405 MHz frequency range, known as Medical Implant Communi-cations System (MICS) band^[33] and its standard, IEEE 802.15.6^[34], appear strategic for communication of ingestible sensors^[35].

Although many standards that are suitable for the ingestible sensors have been released and adopted, further optimization to the communication with more flexibility have been considered. Thus, proprietary protocols with alternative bands^[10],^[36],^[39],^[40] have been proposed for ingestible sensors. Owing to their flexibility, it can further optimize the communication performance and power efficiency. However, they would force to use the additional devices to repeat the communication to off-the-shelf devices, for their compatibility.

Fig. 3 shows the trends in data rate versus communi-cation range of State of The Art (SoTA) wireless modules for ingestible sensors. As aforementioned, higher data rates result in shorter communication ranges. Technological advances are being made to push beyond the 10 Mb/s•m limit.

Fig. 3. Trends in data rate versus communication range of state-of-the-art wireless modules for ingestible sensors.

Implementation of wireless communication for ingestible sensors has several challenges. Monitoring the GI tract using ingestible sensor, which is located deep within the body, results in significant tissue loss during communication. Additionally, the antenna's radiation pattern changes due to body movements and digestive activities, and the antenna's impedance varies with the GI track status, e. g., empty or full stomach. Other wearable or implantable devices located close to the ingestible sensor can interfere with communication with the external hub.

Another challenge is the size of the ingestible sensor. Miniaturization can increase medication compliance and allow easy passage through narrow regions of the GI tract.

Millimeter-scale sensors enable long-term monitoring by attaching the sensor at specific points through gastroscope.

However, as the module size decreases, the capacity of transmit power would be limited, e. g., lower antenna efficiency due to smaller size (to be discussed in Sec. III), which leads to limit to attain sufficient Signal-to-Noise Ratio (SNR) for the desired data rate^[25]:

where $C,\,\,S,\,\,N_{0},$ and BW are channel capacity, signal power, noise spectral density, and bandwidth, respectively. Fig. 4 shows the trends in wireless module area versus data rate of SoTA wireless modules for ingestible sensors.

Fig. 4. Trends in wireless module area versus data rate of state-of-the-art wireless modules for ingestible sensors.

Moreover, miniaturization limits battery capacity, causing difficulties in long term monitoring. For longer sensing while pursuing the miniaturization, energy efficiency operation is required. It is also beneficial to avoid tissue heating due to the high-power consumption of ingestible sensor.

To address these issues, ultra-low-power circuit technologies are necessary. Therefore, miniaturized wireless communication module that can also consume low power is essential for ingestible sensors.

III. Solutions for Wireless System Miniaturization

The wireless system plays a key role in ingestible sensors, transferring the sensor data. To secure communication with the outside world, a link budget analysis shown in Fig. 5 should be conducted, considering the Effective Isotropic Radiated Power (EIRP) of the sensor TX and the path loss (e. g., tissue loss, free-space loss, multipath loss) and ensuring that the received power is sufficiently higher than sensitivity which is the minimum signal strength that can receive the signal. Increasing transmit power or reducing RX sensitivity can increase the link budget but require more power consumption that can reduce the sensor lifetime and occur the tissue heating that would deteriorate the biocompatibility of the sensor.

Fig. 5. Link budget analysis of wireless transceiver.

Increasing the antenna efficiency can also secure the link budget, and thus the communication range. However, it is also unavoidable to increase the antenna size^[41]:

where $\eta ,\,\,\alpha ,$ and $L_{g}$ represent antenna efficiency, attenuation constant and the antenna length, respectively.

To increase the antenna efficiency while maintaining the size, the frequency band of antenna has to be increased as well^[42]:

where $f_{r},\,\,c,\,\,\lambda _{g},\,\,\varepsilon _{eff},$ and $\varepsilon _{r}$ are the resonance frequency, speed of light, wavelength, effective permittivity and combined relative dielectric constant of the substrate, respectively. However, at the higher frequency band, the path loss also increases^[32], which counteracts to link budgeting, and is even more problematic for ingestible sensor communications where EM absorption by tissues is significant in the path loss.

1. Antenna Miniaturization

To miniaturize the antenna while securing the communication distance at the target frequency band, several antenna design techniques have been introduced^[42]. To reduce the size of the antenna with maintaining the resonance frequency, miniaturization techniques are adopted to antenna design. Fig. 6(a) shows a shorting pin^[42]. It is widely used for the miniaturization of implantable antennas. This connects the patch to the ground plane and acts as the ground plane of a monopole antenna that allows to double the electrical size of an antenna. Thus, by adding the shorting pin, the antenna has a similar resonant frequency of an antenna that has twice size without the shorting pin^[42]. The ground plane with slots can provide an additional miniaturization. Since slots added to the ground plane increase capacitance, the resonance frequency is lowered^[45].

Miniaturization is also realized through the pattern of the antenna patch. One of several patterns for miniaturization is a meander line structure, as shown in Fig. 6(b). This structure lengthens the current flow path, so the physical size is decreased with maintaining the electrical length, enabling the operation at the desired frequency. Furthermore, the parasitic capacitance occurring in the gaps between meandering lines makes the resonant frequency move to the lower end of the spectrum^[42]. These antenna miniaturization techniques allow to ensure the same electromagnetic performance even with smaller sizes at the same frequency.

Fig. 6. (a) Antenna structure with shorting pin; (b) Meander line antenna.

2. Antenna Interface Integration

Pattern antennas discussed in Sec. III have narrow-band characteristic and thus their impedance is prone to process variations. Moreover, the antenna impedance varies sensitively depending on the surrounding state, e. g., diet conditions and locations. to adaptively respond to the antenna impedance variation, a Tunable Matching Network (TMN) and matching detector is crucial to ingestible sensors, rather than re-tuning the antenna to a right impedance.

Since the TMN typically comprises passive devices, e. g., inductors, capacitors, it is typically configured with external device^[47] or a separate TMN chip^[48] as it requires large impedances for wide coverage especially at the low frequency. The separate TMN chip is typically implemented in the Silicon-On-Insulator (SOI) process with low-loss substrate^[48], or with multiple inductors and capacitor banks^[49],^[50] which introduces extra loss and silicon area.

To interface from antenna to a transceiver front-end, an antenna switch is typically employed to connect to both TX’s and RX’s TMNs separately, as shown in Fig. 7, but this approach introduces additional loss and area. Instead, the TX/RX shared TMNs directly interface to the antenna can avoid such antenna switch^[51] and further optimize the TMN size. The shared TX/RX Tunable Matching Network (TMN) shares an inductor and uses capacitors to configure the matching network for TX mode and RX mode. Therefore, the TMN is optimized for a specific frequency. If TX and RX use different frequency bands or require a wide bandwidth, the use of the shared TX/RX TMN may be limited. Furthermore, this structure can lead to trade-off in optimizing the performance of each mode. In TX mode, achieving maximum efficiency requires a specific impedance matching condition, but this condition can increase the noise figure in RX mode. Conversely, optimizing the matching for minimal noise in RX mode can result in reduced efficiency in TX mode ^[54].

Fig. 7. Transceiver with Antenna switch and tunable matching network.

On top of the TMN, the impedance matching detection and correction techniques^[48] have also to be adopted to the ingestible sensors, in order for the wireless communication to adapt to its various surrounding state in the body. Traditional techniques^[48],^[55],^[56] use directional coupler^[55],^[56], off-chip tuner^[48] to do the impedance detection, which is bulky at the RF frequencies, e. g., 2.4 GHz and lossy. Recent works^[48],^[50] have demonstrated some techniques to have automatic impedance matching tuning capability to compensate the antenna impedance variations.^[50] uses a two-point amplitude-only detection to extract the impedance mismatch information, and further tuned through an on-chip three stage LC network. It finds the optimum tuning setting through an exhaustive search which can be time consuming, and the work has a relatively small tuning range up to |${\Gamma}$| = 0.3. In^[48], impedance mismatch detection is done by a polar detection which detects both the amplitude and phase information of the impedance seen from the Power Amplitude (PA) output. Based on the polar information, an off-chip tuner with SOI switch is tuned iteratively to reduce the impedance mismatch. With the extra phase information, a successive approximation is used for the optimization, and leads to a faster calibration. In addition, the off-chip SOI tuner covers a large tuning range, up to VSWR of 6 (i. e., |${\Gamma}$| = 0.714). Since^[48] is targeted for cellular application, the power consumption of the detection circuit is relatively high (30 mW). In addition, the separate SOI tuner chip increases the system cost and dimension.

With simpler design approaches, fully integrated on-chip antenna impedance detectors^[32],^[48] have been introduced. The detection technique in^[49] re-uses the balun transformer necessary for the differential PA as a hybrid transformer. As shown in Fig. 8, the hybrid transformer can provide a signal due to impedance mismatch between the real load impedance and an on-chip reference load. With the complex detection by extracting In-phase (I) and Quadrature (Q) information, exact antenna impedance mismatch from the reference load can be detected. Amplitude-only impedance detections^[32],^[57],^[58] can simplify the detection circuity and reduce the power consumption further but it increases the detection time and requires longer settling time for the ingestible sensors to adapt to its surroundings for reliable communication.

Fig. 8. Impedance matching detection with hybrid transformer ^[48].

3. Crystal-less Communication

As discussed in Section II, the required data rate for ingestible sensors are relatively low (~10 Mb/s). Moreover, the communication duration has no need to be long, $\mu s$ to $ms$ scale^[32]. Therefore, to maximize the energy efficiency of communication, the wireless module can be turned off most of time when there is no communication event, so called ``duty-cycling''^[32], as illustrated in Fig. 9.

Fig. 9. Power scenario of duty cycled communication.

In the duty-cycling communication, the wireless module is mostly in the sleeping time while the wakeup timer which wakes up the module in the communication is continuously on. Therefore, the power consumption of the wakeup timer can be dominant in the energy efficiency of the communication. Moreover, due to the timing inaccuracy of the timer, the wireless module should wake up sufficiently in advance of the time when the communication starts (guard time in Fig. 9). Hence, time accuracy of the wakeup timer is also crucial to achieve high energy efficiency of the communication by reducing the operation time of the module.

For this reason, a Crystal Oscillator (XO) has been widely chosen as a wakeup timer, due to its high frequency accuracy and reliability. However, its millimeter-scale size is too bulky to minimize a wireless system for ingestible sensors. To replace the XO, RC based timers^[59] have been introduced as they can be fully integrated on a chip as depicted in Fig. 10 Since the short-term time inaccuracy (jitter) of the timer is averaged out in long-term, in the duty- cycling scenarios, such on-chip timer can provide a decent long- term frequency accuracy. Moreover, RC-based timers can also operate at low power as its frequency is mainly up to the charging-discharging operation, dominant to the RC constant. However, RC-based timers are susceptible to PVT variation; in particular, their sensitivity to temperature (temperature coefficient) is one of the challenging design constraints to overcome^[60]. However, since the body temperature remains relatively constant, they are promising as a wakeup timer for ingestible sensors.

Fig. 10. Relaxation oscillator ^[59].

Frequency stability is another critical constraint to establish the wireless communication, which is guaranteed on the same frequency channel. For example, frequency stabilities in BLE standards^[27] and MICS band^[33] are defined as ${\pm}$41 and ${\pm}$100 ppm, respectively. To meet such accuracy, the XO has been also the inevitable option for a frequency reference^[63]. To minimize the XO size, a Film Bulk Acoustic Resonator (FBAR) is utilized by replacing the quartz crystal^[66]. The FBAR is sub-millimeter sized and secures the high frequency stability (<60 ppm). However, the FBAR requires the additional processing steps and costs for integration.

Alternatively, a network-based frequency synchronization methods have been presented^[32],^[67]]. Instead of external devices, e. g., crystals or FBARs, the network-based frequency synchronization directly utilizes the received signal as a frequency reference.

In general, the network-based frequency synchro-nization technique has two challenges. First, the receiver needs recognize the received signal properly and extract the carrier frequency. Second, the frequency calibration would be susceptible to the carrier frequency drift, as it can be occurred only when the signal is received.

As shown in Fig. 11, tracking loop-based receivers (RXs) extract the carrier frequency from the down-converted Intermediated Frequency (IF) signals and adjust either frequency or phase of the Local Oscillator (LO) accordingly^[32],^[67],^[68]. This behavior can avoid the extra circuitries for frequency or phase detector by detecting frequency or phase from input signal through the mixer. However, this requires the RX to first define the channel to listen on, and thus its LO frequency setting accordingly^[32]. To reduce the energy to initialize the LO frequency, a Phase-Locked Loop (PLL) based carrier frequency extraction with a dedicated reference clock recovery receiver^[69] can be utilized to generate a reference clock from the received signal.

Fig. 11. Tracking loop-based RX.

IV. Future Trends

The utility of ingestible sensors is not limited to GI track monitoring. The ingestible sensor in^[70] demonstrates the potential for vital signal monitoring in the GI track. In addition to the diversity of sensors, integration of sensors in one device, i. e., multi-modality, and a high-resolution sensor are required for an ingestible sensor. These requirements enable more accurate diagnoses and provide new approaches to understanding the complex mechanisms of the GI track. However, due to the high data rate, these sensors should adopt communication at higher frequency band with a wider bandwidth, requiring more power to maintain communication range. For high system integration and stable wireless communications at high frequencies, antenna-RFIC co-design^[71] would be one of the technological trends for ingestible sensors. For example, 3D-shaped and high-Q antennas can serve as inductors in which they can achieve the high volume-efficiency. Low-power and small-area RF frontend designs, e. g., power oscillator (Oscillator + PA)^[71], super-regenerative RX (Oscillator + RX)^[72] and even all-in-one TX/RX frontend^[73] are emerging recently.

To reduce and even eliminate the battery, energy harvesting^[74] can be also embedded to achieve further miniaturization. Battery-free backscattering communi-cation module^[76] are recently emerging in IoT applications. These have great potential for ingestible sensor applications, but ensuring long-term, continuous sensing needs to be addressed.

Also, the ingestible sensor combined with onboard AI accelerator can be a new technological trend^[77]. The AI accelerator enables real-time data processing and decision-making in the sensor and transmits only information data to outside. The sensor also can operate without external controls. AI can drive further miniaturization by reducing the amount of data to and from the outside world and enabling energy-efficient operation with limited battery life.

V. Conclusions

The growth and advancements of ingestible sensor are expected to bring new paradigms to medical diagnostics and treatment, making ingestible sensors a pivotal component of future healthcare systems. Among ingestible sensors, camera pills are popularly utilized in commerce. However, they are limited in the number of diseases they can diagnose. On the other hand, the various biomarkers such as GI fluids, gases, temperature, and pressure enable the accurate and rapid diagnosis of numerous diseases, and research on these ingestible sensors is continuously conducted. This evolution is driven by the continuous development of wireless communication. Since each sensor requires different data rates, appropriate standards and frequency bands is considered to optimize the communication. In addition, ingestible sensors have strict volume constraints and are located deep in the body, where the tissue loss is significant. Therefore, wireless module miniaturization techniques, such as antenna miniaturization, fully integrated antenna interface, and crystal-less communi-cation are required to minimize the module while ensuring reliable communication. It is expected that the miniaturi-zation technique will be continuously evolved. The high integration achieved through antenna-RFIC co-design and battery less operation via backscattering and energy harvesting can advance the miniaturization of ingestible sensors. In addition, the integration of AI into ingestible sensors enables energy-efficient data processing and communication, thereby accommodating increasing sensing data and accelerating the miniaturization by increased system integration and power reduction.