Introduction

Lower urinary tract dysfunction (LUTD) is an umbrella term that includes conditions and diseases related to the bladder and the urethra1. LUTD affected 2.3 billion people worldwide in 2018, which is expected to increase in the future due to the aging population2. Common conditions that cause LUTD include benign prostate hyperplasia (BPH)3 and neurological disorders like multiple sclerosis and spinal cord injury4,5. For BPH-LUTD alone, the global economic burden was estimated to be more than $73 billion in 20196.

Since LUTD revolves around the abnormalities during the filling and emptying of the bladder, bladder volume monitoring is an essential tool to evaluate bladder function7. At present, the gold standard measurement method is bladder catheterization (Supplementary Note 1) 8. Despite its accuracy, it is not recommended for routine examinations due to the high adverse outcomes associated with the invasive procedure9. Hence, clinicians rely on ultrasound imaging of the bladder to estimate the bladder volume, which is a quick and non-invasive method to obtain information such as post-void residual volume of the bladder (Supplementary Note 2). However, this method cannot provide continuous monitoring and necessitates trained personnel with expensive benchtop equipment. Moreover, both of these methods are applicable to patients while they are admitted to the hospital only. Outside the hospital, monitoring relies on the patients themselves to fill urinary diaries and questionnaires to provide information regarding the volume and frequency of micturition10. Although continuous monitoring via implantable sensors does exist at the research level11,12,13,14,15, only patients with complicated or severe symptoms would be suitable for such invasive devices16. Given the importance of bladder volume monitoring and the current status; precise and non-invasive technologies to monitor bladder volume outside the hospital settings are needed17.

To this end, decades of research have led to several commercial wearable devices in which signals from the anterior and posterior walls of the bladder are used to quantify the degree of bladder filling18,19,20. Although such devices might be useful to alert the users when the bladder is full, they are not intended for accurate and continuous bladder volume measurement which is critical for diagnostic and treatment applications21,22,23. In addition, these rigid and non-conformal devices suffer from reliability issues and require ultrasound gel for proper signal transmission, which limits their usage during daily activities24,25. Furthermore, although recent scientific studies reported improved accuracies for continuous monitoring by utilizing phased arrays or multiple elements, these studies either incorporate bulky electronics subsystems or lack such systems totally and require wired connections to benchtop laboratory equipment to conduct the measurements26,27,28,29,30,31. Other reported methods include electrical impedance analysis32,33,34,35 and near-infrared spectroscopy36,37,38. However, these methods could only provide preventative insights, and such indirect measurements would only be suitable for coarse estimations of the bladder volume due to their low specificity, proneness to motion artifacts, and urine-dependent properties39,40. Therefore, none of these devices are suitable for accurate and continuous bladder volume monitoring.

A significant amount of research has been conducted in recent years on wearable ultrasonic transducers, which has opened up a new dimension in the healthcare continuum31,41,42,43,44,45,46,47,48,49. Nevertheless, the majority of the reported devices consist of ultrasonic transducers, which require benchtop electronics and prevent monitoring outside hospital settings. Here, we introduce a wireless and flexible UBVM device with integrated electronics for continuous monitoring with high accuracy. By utilizing multiple transducers for A-mode ultrasonic measurements coupled with a low-power receiver circuitry and a spherical fitting algorithm for bladder volume estimation, the UBVM device enables wireless, non-invasive, and accurate measurement of the bladder volume continuously. The air-backed design of the ultrasonic transducers allows for low-voltage bladder volume measurements of people of all ages. At the same time, a simple and effective time-of-flight measurement algorithm avoids the complex computations commonly associated with ultrasonic imaging and allows the integration of miniaturized electronics to realize a wearable ultrasonic monitoring system. Using a Bluetooth connection from the device to a mobile phone, bladder volume measurements can be monitored continuously. Wireless operation of the system is demonstrated through in vitro and in vivo experiments.

Results and discussion

Design and working principle of the integrated ultrasonic device

The design concept and the operating principle of the UBVM device are illustrated in Fig. 1a. Once placed in the lower abdomen, the UBVM device performs continuous and wireless monitoring of the bladder volume. This is achieved by continuously recording the distance between the anterior and the posterior wall of the bladder from multiple sites using multiple transducers. These measurements are then transferred to a nearby mobile device and utilized in a volume estimation algorithm. The measurement process involves sequentially exciting the transducers, which emit ultrasound waves toward the bladder. In our specific application, these ultrasound waves are reflected from the anterior and the posterior walls of the bladder. Upon returning to the transducers, the waves are converted back into electrical signals, carrying location information of the reflection points from the bladder wall. This information is then transferred to a mobile device via Bluetooth technology. With sequential repetition of such a measurement from multiple transducers at multiple sites, our algorithm estimates a sphere with its surface points corresponding to the reflection points on the bladder surface. The volume of the estimated sphere is then displayed on the custom mobile app we have developed, along with the measurement history (Supplementary Video 1). Further details regarding the operation of the device can be found in the “Methods” section and Supplementary Note 3.

Fig. 1: Design and working principle of the integrated ultrasonic device.
figure 1

a Schematic of the UBVM device placed on the lower abdomen with key functionalities. The device enables continuous bladder volume monitoring by continuously measuring the anterior–posterior (A–P) wall distance of the bladder from multiple sites and processing them to estimate the volume. Within the UBVM hardware, the MCU constantly communicates with the pulser and receiver modules to excite the ultrasonic transducers and record the received echoes from the anterior and posterior walls of the bladder. Received echo timestamps are transferred to a custom-made app for processing and visualization. b Exploded view of the UBVM device, including the printed circuit board (PCB) and layers of the ultrasonic transducers. c Photograph of the ultrasonic transducers.

Full size image

The three main subsystems of the UBVM device are (1) a millimeter-scale conformal ultrasonic transducer module, (2) miniaturized electronics with integrated Bluetooth Low Energy capability, and (3) a custom mobile app that wirelessly communicates with the UBVM hardware and displays the bladder volume measurements (Fig. 1a). In addition to PZT disks and their matching layers, the transducer module consists of a flexible substrate and soft silicone-based encapsulation layer to conform to the human skin (Fig. 1b, c). The module has a footprint of 3 × 3 cm2. The light and compact design facilitates minimal pressure during measurements, as opposed to conventional probes that might cause pain and discomfort due to applied pressure on the skin during the measurements (Supplementary Fig. 1). Moreover, to enable gel-free measurements, a silicone-based couplant was used to eliminate the air in between the transducers and the skin (Supplementary Figs. 2 and 3). From the transducers' point of view, the UBVM device dramatically reduces the required voltage levels by utilizing an air-backed design approach as opposed to conventional medical ultrasonic transducers with backing layers. Such a decrease in power demand not only increases the lifetime of the device with the same battery capacity but also removes the potential of any temperature-induced discomfort, as seen in the thermal images of the UBVM device (Supplementary Fig. 4). The electronics module comprises a microcontroller unit (MCU), along with pulser and receiver circuitry for excitation of the transducers, reception of echo signals and communication with a mobile device. From the electronics design point of view, the UBVM device bypasses power-hungry data acquisition modules by utilizing a 1-bit analog-to-digital conversion scheme as a simple yet efficient time-of-flight measurement strategy (Supplementary Note 3). Such a strategy not only decreases the power consumption of the device but also eliminates additional ICs and peripheral components for the miniaturization of the system. Finally, the custom mobile app communicates with the UBVM hardware every 25 ms for real-time bladder volume monitoring. In addition, the app is capable of communicating with Amazon Web Services (AWS) for potential integration into existing cloud-based patient monitoring services.

Characterization of ultrasonic transducers

Conventional ultrasonic transducers used in medical imaging contain backing layers made up of absorptive materials to dampen the ringing effect, which sacrifices the sensitivity as most of the energy is absorbed by the backing layer (Supplementary Figs. 5 and 6)50. To enable a simple, small footprint and efficient wearable platform, we have opted for an air-backed design with a wrap-around electrode configuration where the former improves the sensitivity of the transducer dramatically and allows measurements with low driving voltages and the latter simplifies the fabrication procedure by enabling electrode access from the same plane (Supplementary Note 4). Figure 2a shows the cross-sectional view of the transducer with structural layers and the air cavity. By using a commercially available copper/polyimide (Cu, 9 μm/PI, 25 μm) flexible laminate, a 40 μm-thick polyethylene terephthalate (PET) encapsulation layer in the back, a thin (100 μm) polydimethylsiloxane (PDMS) encapsulation layer in the front and a silicone-based adhesive, the UBVM transducers provide conformable coupling to the skin and enable gel-free measurements (Supplementary Fig. 2). The robustness of the transducers was tested in a custom-built bending stage, and performance remained unchanged after 500 bending cycles, as seen in Supplementary Fig. 7. In order to compensate for the narrower bandwidth associated with the air-backed design and improve the transducer performance, we have also included a matching layer in front. With a delicate laser etching step, we were able to optimize the matching layer thickness in micron-level precision for optimum transducer performance. (Fig. 2b, Supplementary Notes 5, 6)

Fig. 2: Characterization of the ultrasonic transducers.
figure 2

a Cross-sectional view of the ultrasonic transducers, including each layer and the air-backing. Air-backed design eliminates energy loss due to absorptive backing layers and allows ultrasonic sensing with lower voltages. b Optimization of the matching layer thickness by conductance vs frequency plots. The matching layer was etched until the thickness tf, corresponding to the maximum flat conductance of the transducer. c Pulse-echo response and the derived frequency spectra of a transducer, highlighting the resonance frequency of the transducer. Dashed line represents −6 dB bandwidth. d Impedance and phase angle spectra of a transducer, showing the minimum impedance value at the expected resonance frequency. e Frequency sweep of the transducer captured by a hydrophone. f Echo signal level comparison of the UBVM air-backed transducer to a commercial transducer with a backing layer, showing the enhanced sensitivity of our design. g Mapped pressure field of a UBVM transducer, highlighting the directivity of the transducer. h Cross-sectional field map of the insonation region, parallel to the surface of the transducer.

Full size image

Considering the frequency constraints of the MCU-generated pulses, the deep position of the posterior wall of the bladder, and the frequency-dependent attenuation of ultrasound waves in tissue50, 2 MHz was selected as the operating frequency of the transducer. Figure 2c shows the pulse-echo response of the transducer and its derived frequency spectrum, which shows the resonance frequency of the transducer 2 MHz and reveals a −6 dB bandwidth of 37.9%. Both the electrical impedance and phase angle measurements (Fig. 2d) and the underwater excitation frequency sweep of the transducer (Fig. 2e) also show the 2 MHz resonance frequency of the device, in agreement with each other and the pulse-echo response. To highlight the influence of air-backing, the pulse-echo response of the UBVM device transducer was compared to a commercially available immersion transducer with an operating frequency of 2.25 MHz (Olympus V232-SU) in Fig. 2f. Even though the commercial transducer had a larger active area, the UBVM device transducer demonstrated about 4-fold increase in SNR in comparison to the commercial transducer. As discussed in Supplementary Note 8, such an increase in SNR does not present any adverse effects, as the intensity outputs of the UBVM device are orders of magnitude below the FDA limit. The measured 2D pressure field of the transducer in the longitudinal direction is shown in Fig. 2g, showing the directivity of the transducer as a result of the chosen aspect ratio of the PZT disc. This is further supported by the transverse pressure field scan of the transducer, where the −6 dB contour lines are smaller than the radius of the PZT disc (Fig. 2h and Supplementary Fig. 8).

Electronics design and in-vitro characterization of the system

Figure 3a shows a real photo of the transducers with the fabricated PCB. The system uses six integrated circuits (ICs) to excite the transducers, acquire the received signals, and transmit the recorded timestamps (Fig. 3b). As illustrated in Fig. 3c, the microcontroller unit generates the trigger pulses and multiplexer channel selection codes to address an individual transducer. The multiplexer then transfers the trigger pulses to the chosen channel of the pulser. The pulser generates high voltage pulses for transducer excitation and utilizes its embedded transmit and receive switch to forward the received echo to the receive circuitry. Another multiplexer, controlled by the same selection codes from the MCU, then directs the echo signal, passing through a set of filters and amplifiers, to the comparator. The output of the comparator is sampled by the MCU to acquire the timestamps of the echo signals, which are sent to a mobile phone with a custom app for volume estimation (Supplementary Note 3).

Fig. 3: Electronics design and in-vitro demonstration of the UBVM device.
figure 3

a Photograph of the UBVM device. b Functional block diagram of signal transmission lines of the UBVM device electronics. c Timing diagram of the UBVM electronics. A transducer channel is selected by the MCU and the MUX. Two sets of unipolar pulses with 180° phase difference allow the pulser to generate bipolar high voltage pulses. Received echoes are passed to the receiver circuitry via the internal switch of the pulser. Echoes are amplified, passed through the comparator for echo timestamp detection, and transferred to the mobile phone for processing and visualization. d Schematic drawing of the pulse-echo setup to demonstrate the wireless pulse-echo capability at various frequencies. Echo timestamps displayed on the mobile phone exactly correspond to the raw echo signals on the oscilloscope. e Schematic drawing of the in-vitro setup of the whole system. The test set-up mimics the bladder with flasks in deionized water, where the transducers were connected to the electronics with a 1-m-long flex cable. f Comparison of the actual volume and the measured volume of flasks with various volumes and shapes (n = 10), with a mean relative error of 14.85%.

Full size image

To evaluate the performance of the electronics module, a pulse-echo experiment was conducted in a water tank with commercial transducers. By adapting the clock of our MCU, we were able to excite several commercial ultrasonic transducers at their respective frequencies, and wirelessly transfer the echo timestamps to our custom app in real-time. A steel reflector was positioned at different distances from the transducers to simulate the anterior and posterior wall distances. As depicted in Fig. 3d, the UBVM electronics successfully conducted wireless pulse-echo measurements across the medically relevant frequency range. Once the performance of the UBVM electronics module was validated, the full UBVM device was characterized in vitro in a water tank where the bladder was mimicked by various shaped flasks (Fig. 3e). In addition to spherical-shaped round-bottom flasks, oval-shaped round-bottom flasks, and Erlenmeyer flasks were used to mimic bladder in various shapes and volumes. Measurements were conducted in deionized water, which has a similar acoustic impedance to that of tissue51. As seen in Fig. 3f, the UBVM device estimated the volume of flasks with a mean relative error of 14.85%.

In-vivo validation of continuous bladder volume measurement capabilities

In-vivo measurements were conducted for further demonstration of the UBVM device’s continuous bladder volume monitoring capability. Figure 4a shows a photograph of the transducers on the volunteer’s lower abdomen. The custom mobile app not only processes the measurement data and visualizes them, but also guides the user for optimal placement of the transducers. In an ideal measurement, the transducers should receive eight echo signals in total, four from the anterior and another four from the posterior wall of the bladder. The app analyzes the received signals and alerts the user if the number of echoes received is below five, which is the minimum required for accurate spherical fitting, prompting the user to reposition the transducers.

Fig. 4: In-vivo validation of continuous bladder volume measurement capabilities.
figure 4

a Image of the transducers on the lower abdomen of a volunteer. b Ultrasound image of a volunteer’s bladder with transverse probe orientation. c Ultrasound image of a volunteer’s bladder with longitudinal probe orientation. d Continuous monitoring during a full micturition cycle (red) and the A–P distance measurements of a transducer. Multiple A–P distance measurements from multiple transducers allow spherical volume estimation. e Comparison of the UBVM device measurements with the volumes from conventional ultrasound imaging. A, B, and F represent measurements from the same volunteer at different time points, while C–E and G correspond to measurements from different volunteers (n = 2). f Raw echo signals recorded on an oscilloscope during an in-vivo measurement. Both the anterior and the posterior echo can be seen. g Signal as captured by the MCU after receiver circuitry. These timestamps are used for bladder volume estimations.

Full size image

In order to observe a full micturition cycle, measurements were initiated with an empty bladder, and the transducers were placed with guidance from the mobile app. Volume estimations were recorded every 30 min by opening the custom app for automatic calculation of the bladder volume. Figure 4d shows a full micturition cycle of the volunteer captured by the UBVM device and the A–P distance data from one of the transducers. The increasing bladder volume trend can easily be seen. The UBVM transducers measure the A–P wall distances of the bladder, where multiple distance measurements are used to estimate the volume.

To validate the accuracy of the volume estimates, ultrasound images of the bladder were used to calculate the bladder volumes of five healthy (3 females and 2 males) volunteers (Fig. 4b, c and Supplementary Fig. 9). The ultrasound images were captured by an experienced urologist using a commercial ultrasound imaging system. After the ultrasonographic bladder volume assessments of the participants, the values were compared to the UBVM device estimates. As shown in Fig. 4e, the UBVM device accurately estimated the bladder volumes of the participants, with volumes ranging from as low as 84 mL to as high as 800 mL, demonstrating its performance in various bladder shapes and sizes. This association between the UBVM device results and volume calculation from ultrasound images was further proved by the linear regression analyses, as seen in Supplementary Fig. 10, with an R-squared value of 99.5%.

Raw echo signals and the associated comparator output are shown in Fig. 4f and g. Thanks to the air-backed design with the matching layer, echoes from the anterior and posterior walls of the bladder are clearly captured, even with an excitation voltage as low as 30 V.

Discussion

We have described a flexible and integrated ultrasonic system enabling continuous measurement of the bladder volume. The UBVM device combines a flexible transducer module with an air-backed design and a simple fabrication methodology that enables low driving voltages, an electronics system enabling low-power and miniaturized A-scan measurements with a custom mobile app and data processing strategies for accurate bladder volume measurements.

The UBVM device has two key advantages compared to existing technologies, which either suffer from accuracy issues or require wired connections to benchtop equipment for measurements18,19,20,31 (Supplementary Table 1). Firstly, the UBVM device eliminates the wires current ultrasonic patches require and is fully integrated and wearable. This is of utmost importance to enable unobtrusive measurements and continuous monitoring. Secondly, such full wearability does not carry the burden of unreliable and inaccurate results. By utilizing multiple transducers, the UBVM device offers non-invasive measurements with high accuracy. This superior performance was demonstrated by thorough wearability studies and in-vitro characterizations with a mean relative error of 14.85% among various shaped flasks to mimic various shaped bladders in clinical scenarios. In addition, in-vivo measurements revealed a mean relative error of 11.17% of the patch with various bladder shapes and volumes ranging from 100 to 800 mL.

Future work will focus on increasing the number of transducers while reducing the overall footprint. Although the accuracies of the in-vivo measurements are well within the acceptable range, further improvements in the accuracy are possible with the increased number of coordinates to conduct the volume fitting. In addition, by acquiring at least nine coordinate points, the UBVM device would be able to conduct ellipsoid fitting, which might result in improved estimates and an overall increase in the accuracy of the UBVM device. This increase in the number of transducers could be achieved by increasing the frequency of the transducers, shrinking their size in all three dimensions, and expanding the number of channels of the multiplexers used in the current version of the UBVM electronics. Furthermore, although not the focus of a proof-of-concept study reported here, a more comprehensive in-vivo validation would be completed to evaluate the performance of the device for patients with lower urinary tract dysfunctions and overweight and obese patients. Once these points are addressed, we anticipate the UBVM device to be an invaluable tool in addressing various use cases, such as lower urinary tract symptoms and post-operative urinary retention. In addition, while our focus has been on bladder volume monitoring, extended versions of the system with minor hardware and firmware modifications can open up new dimensions in wearable ultrasound imaging, as shown in recent studies49.

Methods

Design of the UBVM device

Modeling and simulation of the UBVM hardware were performed in LTSpice (Supplementary Figs. 5, 11 and 33). Leach’s equivalent circuit model was used to model the ultrasonic transducers in air backing configuration. An independent voltage source was utilized to excite the transducers, and two ideal operational amplifiers and a comparator were used to model the receive circuitry. To isolate the thickness-mode resonance of the transducers, the aspect ratio of the transducers was chosen to be <0.2 50.

Fabrication of the UBVM device

Bulk PZT disks were purchased from American Piezo (APC-850). To enable access to top and bottom electrodes from the same plane, disks were modified to a custom wrap-around electrode configuration. A commercial picosecond-pulsed laser ablation system (Supplementary Fig. 12); 500 kHz pulse repetition frequency with a divider of 3, 1.5 W power and 1500 mm s−1 laser cutting speed, R4, LPKF Laser & Electronics) was utilized to isolate a small portion of the silver electrode on one face of the PZT disc. Then, the isolated portion of the silver electrode on that face was electrically connected to the electrode on the opposite face of the disc using a low-temperature silver paste (PE827, DuPont).

25 µm-thick flexible laminate with 9 µm copper layer (AC0925, DuPont) was patterned with LPKF U4 to create the circuit pattern of the transducers and cut the laminate for air-backing. Two-component epoxy resin (EPOTEK 301-2FL, 1:0.35) was mixed at 3500 rpm for 20 min using a speed mixer (Hauschild mixer, DAC 150). Al2O3 nanoparticles (290 nm, Nanografi) were added to the mixture for desired acoustic impedance and mixed at 3500 rpm for another 10 min using the same speed mixer. The epoxy-nanoparticle mixture was poured inside a disc-shaped PDMS (Dow Corning, 10:1) mold. Cured matching layer was manually bonded to PZT disks using a fast-curing two-component epoxy (Loctite EA3 422, 1:1). PZT disks with matching layers were manually placed and reflowed with low-temperature solder paste (TS391LT, Chip Quik) (Supplementary Figs. 13 and 14).

To preserve the air-backed design of the transducers, laminate under the disc-shaped PZTs was cut. In analogous fashion, a double-sided tape (Tesa 64621) was machined. PZT-soldered laminate adhered to the bottom encapsulation layer (40 µm-thick PET sheet) using double-sided tape. For the front side encapsulation, a mold was 3D printed (BMF microArch S230) and parylene-c-coated (Plasma Parylene Systems). The laminate with transducers was placed inside the mold. Silicone encapsulant (PDMS, 10:1) was poured and cured at 60 °C for 9 h. To enable gel-free measurements, Silbione 4717 gel (1:1) was screen-printed on the front side of the transducers (Supplementary Figs. 15 and 16). As seen in Supplementary Figs. 2 and 3, Silbione gel offers robust adhesion and very little acoustic attenuation, making it a good choice for wearable ultrasonic applications.

For the fabrication of the UBVM electronics, all components were purchased from Digikey Electronics. Double-sided printed circuit board fabrication started with drilling via holes on copper-clad laminates as per design (LPKF Protomat E44). Drilling step was followed up by pulsed direct current electroplating of copper to connect the top and bottom layers through drilled holes (LPKF Contac S4). Both sides were then patterned using LPKF U4. As shown in Supplementary Fig. 17, main components include a Bluetooth Low Energy integrated microcontroller (STM32WB55RG, STMicroelectronics), a single and a dual channel multiplexer (DG408 and DG409, Maxim Integrated), a high voltage pulser (MAX14808, Maxim Integrated), an operational amplifier (OPA357, Texas Instruments) and a comparator (MAX941, Maxim Integrated) (Supplementary Figs. 18–24).

Matching layer thickness optimization

After identifying the matching layer composition with the desired density and speed of sound (Supplementary Fig. 25), LPKF R4 was utilized (300 kHz pulse repetition frequency with a divider of 2, 7.3 W power and 1400 mm s−1 laser cutting speed) to thin the matching layer into desired thickness. Laser parameters were optimized to remove a micrometer of the matching layer in each pass. After each material removal step, conductance measurements were performed (MFIA Impedance Analyzer, Rohde & Schwarz) in deionized water to find the optimum matching layer thickness. Laser etching followed by an admittance measurement procedure was repeated until the broadest conductance measurement was obtained.

Characterization of the transducers

All electrical and acoustics characterizations were conducted in deionized water. An impedance analyzer (MFIA, Zurich Instruments) was employed to obtain frequency-dependent electrical parameters (electrical impedance, phase, conductance) of the transducer (Supplementary Fig. 32). For the pulse-echo response, an arbitrary waveform generator (33521B Waveform Generator, Keysight) was used to excite the transducer-under-test. The echo signal from a stainless steel reflector was received by an oscilloscope (Picoscope 6000 series) through the use of a diplexer (RDX-6, RITEC Inc.) (Supplementary Fig. 26).

Underwater characterizations were conducted inside a UMS Research system covered with acoustic absorbers (AptFlex F28, Precision Acoustics). The pressure field, sound pressure output, and bandwidth were evaluated with a 0.5 mm needle hydrophone (Precision Acoustics) connected to an oscilloscope (DSOX3024G, Keysight Technologies).

Operation of the UBVM device

During the transmit phase, the microcontroller generates two sets of five 2 MHz pulses amplitude of 3.3 V every 2.5 s (Supplementary Fig. 27). The two sets have a 180° phase difference and are required by the pulser IC for bipolar pulse generation. Through the use of the multiplexer, four individual input channels of the high-voltage pulser are switched at 10-s intervals. Ultrasonic transducers are then excited by bipolar pulses as per the input excitation signals and high voltage DC supply level of the pulser (Supplementary Fig. 28).

During the receive phase, the transmit/receive switch of the MAX14808 is utilized to forward the received signal from the bladder walls to the receive circuitry. The op-amp provides a gain of 10 with 1k and 10k resistors, followed by an RC low-pass filter with a −3 dB cut-off frequency of 5.88 MHz to eliminate the higher order harmonics in the received signal caused by the gain bandwidth product of the amplifiers of the amplifiers. The amplified signal is then fed to the comparator, which is used as a threshold detector. If the received and amplified signal surpasses the comparator threshold (e.g., when an echo is received) the comparator output goes high (5 V). Otherwise, the comparator output is low (0 V) (Supplementary Fig. 11). To block false firings at the output, comparator threshold is dynamically varied through the use of a simple resistive network composed of two 20k resistors that is series between the supply and the ground, the divided voltage then fed back to the comparator’s positive pin over a 60k resistor. The ratio of these resistors determines the amount of the hysteresis. Integrated circuit components used in the transmit and receive phases are powered up by the power management circuit board (PMIC) of the UBVM device (Supplementary Fig. 29). The PMIC board requires a 3.7 V battery connection to provide the desired voltages. These voltage levels include −20, −5, 2.5, 3.3, 5, and 20 V for the components discussed previously. Amongst the transmit, receive, and PMIC electronic circuitry, the PMIC portion of the UBVM device draws the maximum current due to the wide range of required voltages (Supplementary Fig. 30).

The proposed one-bit analog-to-digital conversion allows the use of the input capture mode of the microcontroller. More specifically, a timer channel of the microcontroller is configured to capture the value of a counter at the moment the input signal transitions from one state to another. By connecting the timer channel to the comparator output, time stamps of the echo signals are detected and stored by the microcontroller through the use of direct memory access (DMA). This storage process provides fast capturing of the incoming data without interrupting the work of the microcontroller. Subsequently, this data is transmitted via bluetooth low energy consisting of a generic attribute profile (GATT) server that includes custom services and characteristics. In the context of capturing echo timestamps, one service that contains a data capturing parameter as a characteristic and transfers the data over BLE was used. This characteristic can be accessed by other BLE devices, allowing them to read, write, or be notified. The data sent over BLE is then passed through a custom application on a mobile phone for further processing in Amazon web services (AWS). The phone between the wireless microcontroller and AWS behaves as a gateway for conveying the data using the Internet. The data processing stages involved in AWS allow for the transmission of bladder volume information to the phone, passing through the following services: AWS IoT, AWS Kinesis Data Firehose, AWS S3 Bucket, AWS Lambda, AWS AppSync, AWS Dynamo Table, and finally, the user application. GraphQL language is employed in this pipeline to facilitate interaction between users and the services involved. Users can utilize GraphQL’s capabilities to perform operations such as reading and writing data, as well as connecting their smartphones to the pipeline through an Application Programming Interface (API).

Cloud data processing

The timestamps captured by the microcontroller are processed in the AWS Lambda service using Python. Each timestamp arriving at AWS Lambda triggers the Python code. Therefore, the algorithm continuously registers the timestamps into the data array and checks if the monitoring is complete based on the transducer number information attached to the timestamps. Once the transducer number changes from 4 to 1, the algorithm understands that the monitoring is complete and starts processing the collected data to approximate the bladder volume. First, the timestamp array is masked to determine the number of data points that can be used in the volume calculation. Since a sphere can be defined by a minimum of 4 points, if there is an insufficient number of data points in the array, the algorithm sends an error message to the mobile app. Otherwise, the processing continues with the averaging step. Timestamps for anterior and posterior signals of each transducer are averaged and registered to a new array. Then timestamps in the averaged data array are converted to coordinates assuming a speed of sound of 1480 m/s51. Finally, the coordinate array is fed into the sphere fitting function, which finds the volume of the best-fit sphere using the least-squares method that utilizes the Broyden–Fletcher–Goldfarb–Shanno (BFGS) minimization algorithm. More details regarding the acquired number of coordinate points and the choice of algorithm can be seen in Supplementary Fig. 31.

In vitro characterization

The stainless-steel reflector and round bottom flasks in different volumes were used to evaluate the performance of the developed system in vitro. To test the capabilities of our electronics and firmware, the pulse-echo experiments were conducted at different frequencies. The stainless-steel reflector was attached to a motorized stage of the UMS Research system. Commercial ultrasonic transducers with various operating frequencies (1–2–3–5–8 MHz) were submerged into the UMS Research system and were connected to the developed electronics. The transducers were excited, and the received echoes were captured by our electronics. Captured echo timestamps were then transferred to a mobile phone.

To mimic our in-vivo application, the round bottom flasks were submerged into UMS Research system, along with our transducers (Supplementary Fig. 34). A flex cable was used to connect the transducers to the PCB outside the water tank. Received timestamps were transferred to a mobile phone and converted into coordinate points and fitted to a spherical shape. After the calculations, the volume data of the flask-under test was displayed.

In vivo measurements

Continuous bladder volume measurement was carried out on five healthy volunteers with informed consent and an approved protocol by the Koc University Ethics Committee (2023.006.IRB2.004). Before admission to the clinical study, the volunteers were asked about their health condition and medical history. The criteria for volunteers to be included in the clinical studies were no history of urinary surgery and body-mass index (BMI) between 17 and 25 kg m−2. In addition, volunteers with any kind of electronic implantation were excluded. The BMI of recruited volunteers was between 18 and 23.7 kg m−2. After the consent form signature, volunteers were asked to lay supine for bladder volume measurement.

First, ultrasound imaging of the volunteers was performed in the supine position with a Toshiba Nemio 10 scanner equipped with a 3.5 MHz convex array transducer by an experienced urologist. Ultrasound gel was used for acoustic coupling. The bladder images representing the maximum longitudinal, transverse, and anterior-posterior diameters were recorded. The diameter dimension was determined by manually placing the calipers on the recorded images. Based on the measured dimensions, the volume formula with a correction coefficient of 0.52 was used to calculate the bladder volume52.

Right after the ultrasound imaging was done, the UBVM device was placed on the lower abdomen, just above the pubic bone. A suitable location for the UBVM device was searched manually by the operator with the help of the data processing algorithm feedback. Several measurements were taken using the mobile app, and the data were transferred to a computer. To capture the full micturition cycle, a volunteer was asked to void their bladder entirely. Then, ultrasound imaging and UBVM measurements were repeated every 30 min for 3 h, while the patient was instructed to consume water to speed up the bladder filling processes. After the measurements, a team of clinicians and volunteers were asked to grade the overall experience, as seen in Supplementary Tables 2 and 3.

Statistics and reproducibility

Each attempt at data and sample replication was successful. Specifics of the statistical aspects of the study were discussed in detail in corresponding subsections. No statistical method was used to predetermine the sample size.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.