A postoperative gastrointestinal motility closed-loop management system

Abstract

The application discloses a postoperative gastrointestinal motility closed-loop management system and belongs to the technical field of postoperative gastrointestinal motility nursing, which comprises the following: a wireless distributed sensing node attached to key acupoints on the abdomen for collecting borborygmus, gastrointestinal slow wave, abdominal tension and IMU motion posture; an edge computing control terminal running an adaptive multi-modal fusion algorithm, calculating a gastrointestinal motility index MI(t) by dynamically updating weights through Kalman filtering or variational Bayesian inference; a two-stage safety execution module containing software logic fusing (stopping when the algorithm judges an abnormality) and hardware physical fusing (MOSFET relay forced power-off); and a cloud monitoring platform supporting remote visual monitoring, data tracing, OTA upgrading and adopting AES-256 hardware encryption; the system realizes a four-layer closed-loop architecture from sensing, decision-making, execution and monitoring. The application significantly shortens the postoperative gastrointestinal function recovery time and reduces the incidence of complications through non-invasive multi-modal monitoring and closed-loop feedback.

Description

Technical Field

[0001] This invention belongs to the field of postoperative gastrointestinal motility care technology, and in particular, a postoperative gastrointestinal motility closed-loop management system. Background Technology

[0002] Currently, clinical assessment of postoperative gastrointestinal function mainly relies on subjective inquiries during ward rounds (such as "Have you passed gas today?") and simple physical examinations (such as the frequency of bowel sounds upon auscultation and the degree of abdominal distension upon palpation). These methods are inherently discrete, qualitative, and highly dependent on experience. Even with the use of some auxiliary equipment (such as portable bowel sound recorders or electrogastrography machines), only single-dimensional data can be obtained—for example, a bowel sound recorder can only count the number of sounds per unit time, but cannot distinguish between normal peristaltic sounds and the high-frequency metallic sounds of early intestinal obstruction; although an electrogastrography machine can record slow wave rhythms, its signal is easily interfered with by respiration, body position, or changes in skin impedance, and only reflects the function of the stomach rather than the entire intestine.

[0003] Due to the lack of objective indicators for multimodal fusion, doctors often misjudge patients as "recovering" even when they exhibit early signs of paralytic ileus (such as bowel sounds changing from absent to high-pitched metallic, and persistently increased abdominal tension), continuing to administer prokinetic drugs or electrical stimulation, which actually worsens intestinal ischemia. If this issue is not addressed, postoperative gastrointestinal management will remain in a "blind men and the elephant" stage for a long time, failing to achieve early warning and precise intervention, significantly increasing the risk of catastrophic complications such as intestinal perforation and abdominal infection. Summary of the Invention

[0004] The purpose of this invention is to provide a closed-loop management system for postoperative gastrointestinal motility to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a postoperative gastrointestinal motility closed-loop management system, comprising: Wireless distributed sensor nodes are attached to key acupoints on the abdomen to collect bowel sounds, slow waves of gastrointestinal electrophysiology, abdominal tension, and IMU motion posture. The edge computing control terminal runs an adaptive multimodal fusion algorithm, dynamically updates weights through Kalman filtering or variational Bayesian inference, and calculates the gastrointestinal motility index MI(t). The dual-level safety execution module includes software logic fuse (the algorithm stops when an abnormality is detected) and hardware physical fuse (MOSFET relay forcibly cuts off power). The cloud-based monitoring platform supports remote visual monitoring, data traceability, and OTA upgrades, and uses AES-256 hardware encryption. The system implements a four-layer closed-loop architecture encompassing perception, decision-making, execution, and monitoring. The calculation formula is: ; in, Energy for bowel sounds, For slow wave rhythm regularity, For abdominal tension, This is the IMU motion compensation term.

[0006] In a preferred embodiment of this scheme, the adaptive carrier frequency adjustment module scans the abdominal wall fat thickness using BIA and adaptively sets the carrier frequency based on the abdominal wall fat thickness. The mapping formula is as follows: ; in By amplifying the high impedance difference through natural logarithm, the penetration of the device is optimized for obese patients.

[0007] By employing bioimpedance analysis (BIA) technology, this system can acquire real-time information on abdominal wall fat thickness and dynamically calculate the electrical stimulation carrier frequency based on the ratio of trunk to lower limb resistance. This design addresses the problem of poor penetration in obese patients caused by the fixed frequency (5000Hz) of traditional gastrointestinal electrical stimulation devices (such as VBLOC). Traditional gastric pacemakers can only address individual differences by adjusting current intensity (5-10mA) and pulse width (330μs), but cannot effectively address the differences in electric field distribution caused by varying abdominal wall fat thickness. This technology, through dynamic carrier frequency adjustment, allows the electrical stimulation signal to penetrate abdominal wall fat of different thicknesses, significantly improving the effectiveness of electrical stimulation, especially for obese patients with a BMI > 30 (e.g., in Example 3, the carrier frequency was adjusted to 9.0kHz for a patient with a BMI = 32), avoiding a decrease in treatment efficacy due to insufficient penetration. This design enables the electrical stimulation parameters to adapt to changes in patient body shape, achieving precise treatment.

[0008] In a preferred embodiment of this scheme, the wireless distributed sensing node includes at least two sensing units, which are respectively attached to the Tianshu acupoint and the Zusanli acupoint.

[0009] The standardized table of acupoint parameters is as follows: Tianshu acupoint: Electrical stimulation frequency 2-5Hz sparse-dense wave, airbag pressure 5-10kPa, heat temperature 42-45℃; Zusanli acupoint: electrical stimulation frequency 1-3Hz low-frequency pulse, airbag pressure 3-8kPa, and heat temperature 40-43℃; the system monitors the acupoint impedance in real time through BIA electrodes. If the impedance drops below the threshold, the electrical stimulation intensity is increased or the waveform is switched.

[0010] By combining traditional Chinese medicine theory with modern electrophysiological technology, this system establishes standardized electrical stimulation parameter tables for the Tianshu (ST25) and Zusanli (ST36) acupoints (Tianshu: 2-5Hz sparse-dense wave; Zusanli: 1-3Hz low-frequency pulse). The parameters are dynamically adjusted based on real-time impedance changes at the acupoints monitored by BIA electrodes. This design addresses the problem of existing electrical stimulation therapies lacking unified standards for acupoint selection and parameter settings, relying heavily on subjective experience. In traditional studies, different scholars have shown significant differences in acupoint selection, and parameter settings (such as frequency and intensity) lack unified standards, leading to unstable treatment effects. This technology, through standardized acupoint parameters combined with real-time impedance monitoring (increasing the electrical stimulation intensity or switching the waveform when the impedance drops above a threshold), achieves precision and repeatability in treatment.

[0011] In a preferred embodiment of this scheme, the circuit breaker threshold includes: A skin temperature ≥45°C triggers a hardware power-off. A higher-level circuit breaker is triggered when MI(t) remains below the threshold and the heart rate variability (HRV) is abnormal. The dynamic adjustment model for the circuit breaker threshold is as follows: ; in , The lower the impedance or the higher the trunk impedance, the earlier the fuse threshold is triggered.

[0012] By employing a dynamic formula, the circuit breaker threshold is triggered earlier as MI(t) decreases and the resistance ratio increases, significantly improving the system's sensitivity in identifying dangerous conditions such as intestinal obstruction. For example, when MI(t) remains below the threshold and HRV is abnormal, the circuit breaker threshold decreases by 2°C, causing the system to trigger the circuit breaker earlier and preventing the dangerous condition from worsening. This mechanism makes the safety circuit breaker more accurate and timely, significantly reducing the risk of serious complications such as intestinal perforation.

[0013] In this preferred embodiment, the cloud-based monitoring platform uses AES-256 hardware encryption, supports OTA upgrade signature verification and anti-rollback, and the encryption protocol combines HIPAA compliance with a federated learning framework. Each hospital stores patient data locally and only uploads encrypted MI(t) model increments.

[0014] Hardware-level encryption ensures data security, while a federated learning framework enables local data storage and encrypted model increment sharing across multiple hospitals. This approach protects patient privacy while improving model training efficiency and generalization ability. For example, each hospital stores patient data locally and only uploads encrypted MI(t) model increments, allowing the model to be optimized based on a broader dataset while preventing the leakage of raw data. This design complies with HIPAA compliance requirements and provides a safer and smarter remote monitoring solution for postoperative gastrointestinal motility management.

[0015] In a preferred embodiment of this solution, the system includes a tiered recovery mechanism after circuit breaker failure, which includes: After the hardware fuse fails, the system automatically enters "security monitoring mode"; Continuous monitoring of the patient's condition using low-power sensors; The circuit breaker will be gradually released when multiple parameters (such as MI(t) rise and skin temperature decrease) meet the conditions simultaneously.

[0016] By designing a tiered recovery mechanism after circuit breaker failure (automatic recovery for software-based circuit breakers, manual reset for hardware-based circuit breakers), this system achieves orderly recovery after a safe circuit breaker failure, avoiding the inability to recover after a forced power outage in traditional devices. This design solves the problem of traditional safe circuit breaker mechanisms lacking recovery path planning, which may affect the continuity of treatment. When the circuit breaker condition is released and multiple parameters (MI(t) recovery, skin temperature decrease, abdominal tension reduction) are simultaneously met, the system will gradually release the circuit breaker, rather than simply restoring to the parameters before the power outage.

[0017] In a preferred embodiment of this solution, the system includes a graded alarm system for safety fuse tripping, the system comprising: Level 1 software circuit breaker triggers an alarm at the nurse station. A secondary hardware fuse failure triggers an audible and visual alarm. The system establishes a correspondence between alarm levels and circuit breaker levels, and supports remote visual monitoring, data traceability, and OTA upgrades.

[0018] In a preferred embodiment, the system includes an adaptive electrical stimulation waveform unit driven by bowel phonation spectrum characteristics, the unit comprising: By analyzing bowel sounds using FFT, metallic frequency and rhythm disorder bands were extracted. Establish a mapping table between spectral characteristics and electrical stimulation parameters; Edge computing terminals analyze the spectrum in real time, dynamically adjust the waveform, and record therapeutic feedback.

[0019] In a preferred embodiment, the system includes a closed-loop control unit for joint decision-making based on traditional Chinese medicine syndrome differentiation and physiological parameters, the unit comprising: Combining data from the four diagnostic methods of traditional Chinese medicine with quantitative indicators of postoperative gastrointestinal motility; Construct a joint scoring model of "syndrome type-physiological parameters"; The edge computing terminal updates the certificate score every hour and adjusts the electrical, mechanical and thermal parameters accordingly.

[0020] In a preferred embodiment, the system further includes a flexible piezoelectric sensor array integration unit, which comprises: Deploy a PVDF nanofiber piezoelectric sensor array in the abdominal sensing node; Simultaneous acquisition of acoustic vibrations (bowel sounds) and abdominal tension; Real-time calibration is performed using an edge computing terminal to offset changes in electrode contact impedance.

[0021] Compared with the prior art, the technical effects and advantages of the present invention are as follows: This postoperative gastrointestinal motility closed-loop management system, through a four-layer closed-loop architecture of "sensing-decision-execution-monitoring," enables real-time monitoring, dynamic assessment, and precise intervention of postoperative gastrointestinal motility disorders, forming a complete diagnosis and treatment closed loop. At the sensing layer, non-invasive, patch-attached multimodal sensing nodes are used instead of traditional invasive implantable electrodes, avoiding the trauma risks and postoperative complications associated with laparoscopic electrode implantation required by devices such as the VBLOC system.

[0022] These sensor nodes simultaneously collect data on bowel sounds, slow wave electrocoagulation (SWC) of the gastrointestinal tract, abdominal tension, and movement posture, providing comprehensive and multi-dimensional physiological information for the decision-making level. At the decision-making level, the edge computing terminal calculates the gastrointestinal motility index MI(t) through a dynamic multimodal fusion algorithm, rather than relying on a single parameter or subjective assessment (such as traditional radionuclide scanning or contrast imaging methods), achieving an objective quantitative assessment of gastrointestinal motility status. At the execution level, a dual-level safety circuit breaker mechanism (software logic circuit breaker + hardware physical circuit breaker) ensures treatment safety, immediately stopping stimulation when intestinal obstruction characteristics or abnormal vital signs are detected, preventing complications caused by traditional devices (such as gastric pacemakers) failing to automatically identify dangerous conditions and continuing stimulation. At the monitoring level, the cloud platform employs AES-256 hardware encryption and a federated learning framework to achieve secure data sharing and remote centralized management, solving the problem of medical data silos.

[0023] Compared to existing technologies, this technology achieves precise, personalized, and intelligent postoperative gastrointestinal motility management, significantly improving treatment efficacy and safety. Traditional methods, such as the BVBLOC system, can block vagal nerve signals through high-frequency electrical pulses (5000Hz), but lack a real-time feedback mechanism and cannot dynamically adjust stimulation parameters based on individual patient differences (such as abdominal fat thickness), and carry risks of postoperative complications such as infection and wire breakage. This technology, through non-invasive multimodal monitoring and closed-loop feedback, significantly shortens postoperative gastrointestinal function recovery time and reduces the incidence of complications. Attached Figure Description

[0024] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1This is a diagram illustrating the architecture of a closed-loop management system for postoperative gastrointestinal motility according to the present invention. Figure 2 This is a flowchart of a closed-loop management system for postoperative gastrointestinal motility according to the present invention. Detailed Implementation

[0026] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described in order to avoid obscuring the invention.

[0027] This embodiment provides, for example Figures 1 to 2 The postoperative gastrointestinal motility closed-loop management system shown adopts a four-layer closed-loop architecture of "perception-decision-execution-monitoring". The layers are connected through a wireless communication protocol to realize closed-loop control of data acquisition, processing, decision-making and execution.

[0028] Perception Layer: Wireless distributed sensor nodes are attached to key acupoints on the abdomen, including Tianshu (ST25) and Zusanli (ST36). Each sensor node contains multiple sensors to collect bowel sounds, slow-wave electrical activity in the gastrointestinal tract, abdominal tension, and IMU motion posture data. The sensor nodes communicate with the edge computing control terminal through a BLE Mesh network, forming a multi-node network.

[0029] Decision-making layer: Edge computing control terminals are deployed at the patient's bedside or in wearable hosts, running adaptive multimodal fusion algorithms. They dynamically update weights through Kalman filtering or variational Bayesian inference, calculate the gastrointestinal motility index MI(t), and generate personalized treatment strategies. The terminals possess sufficient computing power to process multimodal data and make decisions in real time.

[0030] Execution Layer: The dual-level safety execution module is responsible for delivering outputs such as electrical stimulation, airbag compression, and thermotherapy, and features both software logic circuit breakers and hardware physical circuit breakers. The module employs a low-power design, allowing for extended bedside operation.

[0031] Monitoring Layer: The cloud-based monitoring platform communicates with the edge computing control terminal via the MQTT protocol, supporting remote visual monitoring, data traceability, and OTA upgrades. It also employs AES-256 hardware encryption to ensure data security. The platform can connect to multiple patient terminals simultaneously for centralized management.

[0032] The system workflow is as follows: sensor nodes collect multimodal data → edge terminals calculate MI(t) → treatment strategies are generated based on the MI(t) state → the execution module executes the output → remote monitoring is performed on the cloud platform. The entire process forms a closed loop, ensuring the real-time nature and safety of the treatment.

[0033] In this embodiment, each wireless distributed sensing node includes the following core components: Acoustic sensor: A MEMS microphone with a sampling frequency of 44.1kHz is used to collect bowel sound signals. The microphone is encapsulated in a waterproof and dustproof housing, and low-frequency response is enhanced by a piezoelectric ceramic plate.

[0034] Electrophysiological sensor: Employs an Ag / AgCl electrode array with a 2cm electrode spacing for acquiring slow-wave electrical signals from the gastrointestinal tract. The electrode surfaces are silver-plated, with an impedance of <5kΩ to ensure signal quality.

[0035] Abdominal tension sensor: A PVDF nanofiber piezoelectric sensor array is used, with a sensor array size of 3×3cm and a sensitivity of >10pF / kPa, to measure changes in abdominal tension.

[0036] IMU module: Employs a six-axis accelerometer / gyroscope with a sampling frequency of 100Hz, used to collect patient motion posture data and compensate for motion interference from acoustic sensors.

[0037] BIA electrode: A four-electrode method is used, with two transmitting electrodes and two receiving electrodes, with an electrode spacing of 3 cm, to scan the thickness and impedance distribution of abdominal wall fat.

[0038] Microcontroller: Uses ESP32-C6 chip, supports BLE 5.2 Mesh communication and 2.4GHz Wi-Fi dual-mode connection, and has sufficient processing power to run data preprocessing algorithms.

[0039] Power module: Powered by button batteries, with solar charging panels to extend battery life; battery capacity can support continuous operation for 72 hours.

[0040] In this embodiment, the acupoints are selected based on the following traditional Chinese medicine theories and modern research: the sensor nodes are attached to the Tianshu (ST25) and Zusanli (ST36) acupoints. Tianshu acupoint: Located 2 cun lateral to the navel, it is the Mu point of the large intestine. Traditional Chinese medicine believes that stimulating this acupoint can regulate intestinal motility. Modern research shows that electroacupuncture at Tianshu acupoint can promote colorectal motility by activating the enteric nervous system. Its effect is related to stimulating the enteric nervous system and affecting the release of neurotransmitters such as serotonin and substance P.

[0041] Zusanli (ST36): Located 3 cun below the knee, one finger-width lateral to the anterior crest of the tibia, it is the lower He-Sea point of the stomach. Traditional Chinese medicine believes that stimulating this point can regulate and tonify the spleen and stomach, and harmonize the Qi of the internal organs. Modern research shows that electroacupuncture at Zusanli can improve intestinal motility disorders in rats, increase gastric discharge frequency, and promote gastric motility by activating sympathetic or vagus nerve pathways.

[0042] In this embodiment, the multimodal data acquisition method includes: bowel sound acquisition: an acoustic sensor acquires abdominal acoustic signals, which are then filtered by a low-pass filter (cutoff frequency 200Hz) and a high-pass filter (cutoff frequency 10Hz) to remove environmental noise and baseline drift. The signal is then processed by RMS to obtain the bowel sound energy, measured in dB.

[0043] Gastrointestinal slow-wave acquisition: Electrophysiological sensors acquire gastrointestinal electrical activity signals, and slow-wave rhythms are extracted through bandpass filtering (0.5-5Hz). Slow-wave regularity is calculated through power spectral analysis; the higher the regularity, the larger the value.

[0044] Abdominal tension acquisition: The abdominal tension sensor converts mechanical force into an electrical signal, which is then amplified and filtered before being converted by an ADC to obtain the abdominal tension, measured in kPa.

[0045] IMU Motion Acquisition: The IMU module acquires acceleration and angular velocity data, and fuses the accelerometer and gyroscope data through Kalman filtering to obtain the patient's motion posture, which is used to compensate for motion interference from the acoustic sensors.

[0046] BIA scan: BIA electrodes measure resistance values ​​in two frequency bands, 5kHz and 50kHz, and calculate the resistance ratio of the torso to the lower limbs (50kHz) / (5kHz) for dynamic adjustment of the carrier frequency.

[0047] In this embodiment, the wireless communication mechanism is as follows: the sensor node communicates with the edge computing control terminal through a BLE Mesh network, using an intra-group address packet control protocol to reduce the address bits required for each device in the same group. Compared with the traditional method using source addresses, the time required to transmit the same data packet can be reduced by 23.5%. The network supports nanosecond-level time synchronization, ensuring the synchronization of data across multiple nodes.

[0048] In this embodiment, the edge computing control terminal uses a Xilinx Zynq-7000 series chip, which includes an ARM Cortex-A9 processor and an FPGA programmable logic unit. The terminal hardware configuration is as follows: Processor: ARM Cortex-A9 dual-core processor, clock speed 667MHz, responsible for running the operating system and algorithm logic.

[0049] FPGA: 700MHz clock frequency, containing programmable logic units, used for hardware acceleration of the Kalman filter algorithm.

[0050] Storage: 1GB DDR3 RAM, 16GB eMMC flash memory for storing algorithm parameters and historical data.

[0051] Communication module: Supports BLE 5.2 Mesh and Wi-Fi dual-mode connectivity for communication with sensor nodes and cloud platforms.

[0052] Encryption module: Supports AES-256 hardware encryption and connects to the processor via the I²C interface.

[0053] In this embodiment, the MI(t) calculation implementation includes: implementing the Extended Kalman Filter (EKF) algorithm in the FPGA to handle the nonlinear system state estimation problem.

[0054] In this embodiment, the dynamic weight update mechanism is as follows: The weights are dynamically adjusted based on the following factors: α(t), β(t), γ(t), δ(t): Signal-to-noise ratio (SNR): The lower the SNR, the smaller the corresponding weight.

[0055] BMI value: The higher the BMI, the worse the abdominal penetration and the smaller the weight of the electrophysiological sensor.

[0056] Electrode bonding quality: When the impedance is >5kΩ, the corresponding sensor weight decreases.

[0057] In this embodiment, the electrical stimulation carrier frequency is dynamically adjusted according to the mapping formula. Based on the acupoint parameter standardization table and combined with real-time monitoring of acupoint impedance changes by the BIA electrode, the electrical stimulation parameters are dynamically adjusted. When the impedance drops beyond a threshold, the electrical stimulation intensity is increased or the waveform is switched. The fuse threshold is dynamically adjusted according to the formula of the dynamic adjustment model for the fuse threshold.

[0058] In this embodiment, a two-stage safety circuit breaker mechanism is employed. Software circuit breaker logic: The edge computing control terminal monitors the following parameters in real time to trigger a software circuit breaker: Intestinal obstruction characteristics: (1) By analyzing the bowel sound spectrum using FFT, the metallic sound frequency (>1000Hz) and the rhythm disorder band (sudden increase in energy in the low frequency band) were extracted. (2) When the metallic sound energy accounts for >30% and lasts for >10 seconds, it is judged as a risk of intestinal obstruction.

[0059] Changes in abdominal tension: Calculate the abdominal tension gradient ΔT_tension / Δt. When the gradient is >5 kPa / s and lasts for >5 seconds, it is considered to be at risk of intestinal obstruction.

[0060] Abnormal vital signs: Monitor SpO2 and heart rate changes; trigger circuit breaker when SpO2 < 90% or heart rate drops by more than 20%.

[0061] Software circuit breaker process: (1) When intestinal obstruction features or abnormal vital signs are detected, the treatment signal output should be stopped immediately.

[0062] (2) Send a circuit breaker event notification to the cloud platform, including information such as the reason for the circuit breaker, the time, and the patient ID.

[0063] (3) Before triggering the hardware circuit breaker, the software circuit breaker should be triggered first to allow medical staff to intervene.

[0064] In this embodiment, the MOSFET relay selection is as follows: a normally closed MOSFET relay is used, which supports a 60V load voltage and has an on-resistance of <10Ω, ensuring that the power supply circuit is automatically cut off in the event of a fault.

[0065] Drive circuit design: A gate driver chip is connected to a MOSFET relay, and the control signal comes from the processor's GPIO pin. The circuit topology is as follows: the gate is connected to the output pin of the driver chip, the source is connected to the power input of the treatment circuit, and the drain is connected to the power output of the treatment circuit.

[0066] Watchdog circuit design: A hardware watchdog chip is used, which works in conjunction with the processor's independent watchdog. The watchdog timeout is set to 2 seconds.

[0067] In this embodiment, the two-stage circuit breaker linkage logic is as follows: Software circuit breaker trigger conditions: (1) MI(t) characteristics meet “high frequency metallic sound (>1000Hz) + high tension (>30kPa)”; (2) SpO2 <90%; (3) heart rate drops by more than 20%.

[0068] Hardware fuse trigger conditions: (1) MCU crash / watchdog overflow; (2) current overload (>2A); (3) skin temperature ≥45℃.

[0069] Circuit breaker response method: Software circuit breaker: Stops the output of treatment signals and sends an alarm to the nurse station.

[0070] Hardware fuse: Physically cuts off the power supply circuit (MOSFET relay forces power off), triggering an audible and visual alarm.

[0071] Recovery mechanism after fuse failure: After the fuse fails, the system automatically enters "safety monitoring mode," continuously monitoring the patient's condition through low-power sensors. The fuse is gradually deactivated when the fuse failure conditions are lifted and multiple parameters (MI(t) recovery, skin temperature decrease, abdominal tension reduction) are simultaneously met. Manual reset is required for hardware fuse failure to ensure safety.

[0072] In this embodiment, the cloud-based monitoring platform: Communication Protocol: The cloud-based monitoring platform and the edge computing control terminal communicate using the MQTT over BLE Mesh protocol, specifically as follows: BLE Mesh Networking: Edge terminals act as proxy nodes in the BLE Mesh network, aggregating data from the abdominal sensor nodes. The network employs a nanosecond-level time synchronization mechanism to ensure the synchronization of data across multiple nodes.

[0073] MQTT protocol encapsulation: Data collected by BLE Mesh is encapsulated into MQTT messages via a Wi-Fi module and uploaded to the cloud platform. The message topic design is as follows: Monitoring data: / Hospital ID / Ward ID / Patient ID / MI(t); Circuit Breaker Event: / Hospital ID / Ward ID / Patient ID / Circuit Breaker Event; OTA upgrade command: / hospital ID / ward ID / patient ID / ota command.

[0074] QoS level settings: Set different service quality levels according to the importance of the data: Monitoring data: QoS 2 (ensuring reliable transmission); Circuit breaker event: QoS 1 (reduce transmission latency); OTA upgrade command: QoS 2 (ensure reliable transmission).

[0075] In this embodiment, the encryption mechanism is as follows: The cloud monitoring platform uses AES-256 hardware encryption, specifically implemented as follows: Data segmentation: The collected multimodal data is segmented into 16-byte blocks.

[0076] Key management: The master key is stored in the hardware security module, and each patient is assigned an independent data encryption key.

[0077] Encryption mode: AES-GCM mode is used to generate authentication tags to ensure data integrity and authenticity.

[0078] HIPAA Compliance: Implementing Field-Level Encryption Sensitive fields (MI(t), circuit breaker events) are transmitted in encrypted form.

[0079] Non-sensitive fields (timestamp, patient ID) are transmitted in plaintext.

[0080] Key rotation cycle: Keys are automatically rotated every 90 days, and old keys are archived to the hardware security module.

[0081] In this embodiment, the OTA upgrade process is as follows: The cloud monitoring platform supports OTA upgrades, and the specific implementation is as follows: Upgrade package segmented transmission: The upgrade package is divided into 1MB segments and transmitted to the edge terminal segment by segment via the MQTT protocol.

[0082] Signature verification: The edge terminal uses an RSA public key to verify the upgrade package signature, ensuring that the upgrade source is trustworthy.

[0083] Version verification: The upgrade package must contain a version number, and the new version number must be greater than the current version number.

[0084] Dual-partition storage: Employs an A / B partition storage mechanism, with upgrade packages written to the spare partition.

[0085] Secure Boot: When the bootloader starts, it checks the partition status and signature, and prioritizes the use of higher version partitions to prevent upgrade rollback.

[0086] In this embodiment, the initialization process is as follows: (1) Attach the sensing nodes to the Tianshu and Zusanli acupoints to ensure good contact between the electrodes and the skin.

[0087] (2) The resistance values ​​were measured in two frequency bands, 5kHz and 50kHz, using the four-electrode method, and the resistance ratio of the trunk to the lower limbs was calculated.

[0088] (3) Based on the BIA scan results, the electrical stimulation carrier frequency is dynamically adjusted using a formula.

[0089] (4) Perform sensor calibration, including baseline calibration of acoustic sensors, impedance calibration of electrophysiological sensors, and zero-point calibration of abdominal tension sensors.

[0090] In this embodiment, the monitoring process is as follows: (1) Multimodal data acquisition: Sensor nodes acquire bowel sounds, slow waves of gastrointestinal electrophysiology, abdominal tension and IMU motion posture data.

[0091] (2) Data preprocessing: Acoustic data: Low-pass filtering (cutoff frequency 200Hz) and high-pass filtering (cutoff frequency 10Hz) remove ambient noise and baseline drift.

[0092] Electrophysiological data: slow wave rhythms were extracted using bandpass filtering (0.5-5Hz).

[0093] Abdominal tension data: After amplification and filtering, the values ​​are obtained through ADC conversion.

[0094] IMU data: Kalman filter fusion of accelerometer and gyroscope data.

[0095] (3) MI(t) calculation: The gastrointestinal motility index is calculated by the Kalman filter algorithm.

[0096] In this embodiment, the decision-making process is as follows: (1) MI(t) analysis: The trend of MI(t) is analyzed at the edge terminal to determine the gastrointestinal motility status.

[0097] (2) Treatment strategy generation: If MI(t) is low and there are no signs of intestinal obstruction: activate "wake-up mode" (low-frequency sparse-dense wave + balloon compression).

[0098] If slow wave recovery is detected: switch to "boost mode" (synchronized pulse stimulation).

[0099] (3) Parameter adjustment: Electromyographic tolerance: Automatically increase intensity by 10% or switch waveforms.

[0100] Changes in body position: When lying flat, stop strong stimulation of the stomach (to prevent aspiration).

[0101] In this embodiment, the execution flow is as follows: (1) An electrical stimulation signal with a carrier frequency f_carry is generated by a DAC chip and output to the acupoints on the abdomen.

[0102] (2) Adjust the airbag pressure to the target value through the air pump control circuit (according to the standardized table of acupoint parameters).

[0103] (3) Adjust the temperature to the target value (according to the standardized table of acupoint parameters) by controlling the heating element circuit.

[0104] In this embodiment, the monitoring process is as follows: (1) The edge terminal uploads MI(t) data and circuit breaker events to the cloud monitoring platform.

[0105] (2) Medical staff monitor the patient’s status in real time through the cloud platform, including MI(t) trend chart, circuit breaker event records, etc.

[0106] (3) The cloud platform pushes the upgrade package to the edge terminal to complete the firmware update.

[0107] In this embodiment, the recovery process after the circuit breaker is triggered is as follows: After the circuit breaker trips, the system automatically enters "safety monitoring mode".

[0108] The patient's condition is continuously monitored using low-power sensors, including MI(t) recovery, skin temperature decrease, and abdominal tension reduction.

[0109] When multiple parameters simultaneously meet the conditions, the circuit breaker is gradually released: Software circuit breaker: Automatic recovery, no manual intervention required. Hardware circuit breaker: Requires manual reset by medical personnel.

[0110] In this embodiment, the bowel sound feature extraction is as follows: (1) Bowel sound signals are acquired using a sampling rate of 44.1 kHz. (2) The signal is segmented using a Hamming window function, with each segment consisting of 1024 sampling points. (3) The bowel sound spectrum is analyzed using FFT to extract the following features: Metallic frequency: The proportion of high-frequency components with energy above 1000Hz.

[0111] Rhythm disorder band: Low frequency band (<100Hz) energy surge.

[0112] Mapping of spectral characteristics to electrical stimulation parameters: Metallic audio frequency >1000Hz: Triggers density enhancement (frequency increased to 5Hz, duty cycle increased to 60%), simulating "purging" to clear intestinal gas.

[0113] Low-frequency energy surge: Switching synchronous pulse stimulation (frequency reduced to 1Hz, pulse width extended to 500μs) simulates "complementation" to promote peristalsis.

[0114] Dynamic adjustment mechanism: Edge computing terminals analyze the spectral characteristics of bowel sounds in real time.

[0115] The electrical stimulation waveform parameters are dynamically adjusted based on changes in characteristics.

[0116] The system records therapeutic feedback to optimize the mapping relationship between spectral characteristics and electrical stimulation parameters. By using bowel sound spectral characteristics to guide the optimization of electrical stimulation waveforms, it achieves digital simulation of "purging" and "tonifying" methods, improving the precision and personalization of treatment.

[0117] By employing FFT analysis to analyze the bowel sound spectrum, extracting metallic frequencies and rhythmic disorder bands, and establishing a mapping table between spectral characteristics and electrical stimulation parameters, this system achieves dynamic adjustment of the electrical stimulation waveform based on bowel sound characteristics. This design solves the problem that traditional electrical stimulation devices, which rely on fixed waveforms (such as sparse-dense waves and low-frequency pulses), cannot accurately match pathological states. For example, when the energy proportion of high-frequency metallic sounds (>1000Hz) in bowel sounds is detected to be >40%, the system triggers sparse-dense wave enhancement (frequency increased to 5Hz, duty cycle increased to 60%), simulating a "purging" method to clear intestinal gas; when the low-frequency energy suddenly increases, the system switches to synchronous pulse stimulation, simulating a "tonifying" method to promote peristalsis. This adaptive waveform adjustment based on acoustic characteristics allows electrical stimulation to more accurately match the patient's actual gastrointestinal state, significantly improving treatment efficacy and patient comfort.

[0118] In this embodiment, the closed-loop control unit for joint decision-making based on TCM syndrome differentiation and physiological parameters is as follows: Data collection for the four diagnostic methods of Traditional Chinese Medicine: Tongue diagnosis: The tongue image is captured by a camera, and features such as RGB color values ​​and tongue coating thickness are extracted.

[0119] Face consultation: Facial images are captured through a camera, and features such as facial color and expression are extracted.

[0120] Pulse diagnosis: The pulse waveform is acquired through a piezoelectric / piezoresistive sensor, and features such as frequency and amplitude are extracted.

[0121] Joint scoring model construction: Combining TCM syndrome differentiation theory (such as Qi stagnation type and spleen deficiency type) with postoperative gastrointestinal motility quantitative indicators.

[0122] Constructing a joint scoring model of "syndrome type-physiological parameters": Qi stagnation type: purple tongue (RGB value threshold), wiry pulse (frequency characteristics), MI(t) < 0.3.

[0123] Spleen deficiency type: sallow complexion, weak pulse (amplitude characteristics), MI(t) < 0.2 and abdominal fat thickness > 30 mm.

[0124] In this embodiment, the dynamic adjustment mechanism is as follows: (1) The edge computing terminal updates the certificate type score every hour.

[0125] (2) Adjust the electrical-thermal parameters in conjunction with the scoring results: For Qi stagnation type: increase the intensity of electrical stimulation of Zusanli (3-5Hz), increase the pressure of Tianshu acupoint (8-10kPa), and maintain the temperature of Tianshu acupoint (42℃).

[0126] For spleen deficiency type: prolong the warming time of Tianshu acupoint (maintain 45℃ for 15 minutes), reduce the frequency of electrical stimulation (1-2Hz), and reduce the pressure of the air sac (5kPa).

[0127] (3) Record efficacy feedback to optimize the combined scoring model.

[0128] By combining data from the four diagnostic methods of Traditional Chinese Medicine (TCM) (tongue diagnosis, facial diagnosis, and pulse diagnosis) with postoperative gastrointestinal motility quantitative indicators, a joint scoring model of "syndrome type-physiological parameters" is constructed. The system updates the syndrome type score hourly to adjust treatment parameters accordingly, enabling the generation of personalized treatment strategies that integrate TCM and Western medicine. This design addresses the problem of traditional treatments failing to integrate TCM syndrome differentiation with physiological parameters and relying solely on single indicators or experience to adjust parameters. For example, when a patient is diagnosed with "qi stagnation type" (purple tongue, wiry pulse, MI(t) < 0.3), the system will increase the intensity of electrical stimulation at Zusanli (ST36) (3-5Hz) and increase the pressure at Tianshu (ST25) (8-10kPa). When a patient is diagnosed with "spleen deficiency type" (sallow complexion, weak pulse, MI(t) < 0.2, and abdominal fat thickness > 30mm), the system will prolong the warming time at Tianshu (ST25) and decrease the electrical stimulation frequency (1-2Hz). This combined TCM and Western medicine decision-making mechanism makes treatment strategies more comprehensive and precise, significantly improving the efficiency and quality of postoperative gastrointestinal motility recovery.

[0129] In this embodiment, the flexible piezoelectric sensor array integration unit is: (1) Sensor array fabrication: A PVDF nanofiber piezoelectric sensor array was fabricated using a solution method. The piezoelectric coefficient (d) was improved by a corona polarization process. 33 (Increased from 8.6 pC / N to 32.9 pC / N). A damping layer was added to suppress high-frequency oscillations and improve signal stability.

[0130] (2) Synchronous acquisition mechanism: Acoustic vibration (bowel sounds) and abdominal tension are acquired simultaneously. An alternating acquisition strategy is adopted: acoustic vibration and tension data are acquired alternately 10 times per second. Multimodal data are synchronized through timestamps to ensure acquisition accuracy.

[0131] (3) Real-time calibration algorithm: Real-time calibration is performed using an edge computing terminal. The sensor voltage response curve is calibrated using an oscilloscope and a high impedance analyzer. The compensation parameters are dynamically adjusted according to the changes in electrode contact impedance to counteract the effects of contact impedance changes.

[0132] The flexible piezoelectric sensor array enables the synchronous acquisition of acoustic vibration and abdominal tension, and the real-time calibration algorithm improves the accuracy and stability of data acquisition, providing more comprehensive technical support for gastrointestinal motility monitoring.

[0133] By deploying a PVDF nanofiber piezoelectric sensor array in the abdominal sensing node, simultaneously acquiring acoustic vibrations (intestinal sounds) and abdominal tension, and utilizing an edge computing terminal for real-time calibration to offset changes in electrode contact impedance, this system achieves high-precision and high-stability gastrointestinal motility monitoring. This design solves the problems of poor data synchronization and susceptibility to changes in contact impedance in traditional discrete sensors. This technology significantly improves the synchronization and reliability of multimodal data through an alternating acquisition strategy (acoustic vibration and tension data are acquired alternately 10 times per second) and a real-time calibration algorithm.

[0134] In this embodiment, a graded alarm system for safety fuse failure is implemented. (1) Alarm hierarchy design: Level 1 software circuit breaker: Push an alarm to the nurse station to notify medical staff to intervene.

[0135] Level 2 hardware fuse failure: Triggers an audible and visual alarm, alerting medical staff to take immediate action.

[0136] Alarm classification and circuit breaker level correspondence: software circuit breaker corresponds to level 1 alarm, and hardware circuit breaker corresponds to level 2 alarm.

[0137] (2) Alarm triggering conditions: Level 1 software circuit breaker: MI(t) remains below the threshold and HRV is abnormal.

[0138] Level 2 hardware fuse failure: Skin temperature ≥ 45℃ or current overload > 2A.

[0139] (3) Alarm response mechanism: After receiving a Level 1 alarm, the nursing station can check the patient's status through the cloud platform and decide whether on-site intervention is needed.

[0140] Upon receiving a Level 2 alarm, medical staff must immediately go to the patient's bedside to check the cause of the circuit breaker failure and take appropriate action.

[0141] The safety circuit breaker graded alarm system enables graded responses to circuit breaker events, improving the response efficiency of medical staff and ensuring patient safety.

[0142] By designing a hierarchical alarm system with a first-level software-triggered alarm pushing to the nurse station and a second-level hardware-triggered alarm triggering audible and visual alarms, this system achieves hierarchical response to circuit breaker events, improving the response efficiency of medical staff. This design solves the problems of unclear hierarchical classification and low response efficiency in traditional alarm systems. This technology directly links the circuit breaker level with the alarm method: a first-level alarm sends a notification to the nurse station, and a second-level alarm triggers an audible and visual alarm, notifying medical staff to handle the situation immediately.

[0143] In this embodiment, a tiered recovery mechanism is implemented after the circuit breaker is triggered: (1) Safety monitoring mode: After the circuit breaker trips, the system automatically enters "safety monitoring mode".

[0144] The patient's condition is continuously monitored using low-power sensors, including key parameters such as MI(t), skin temperature, and abdominal tension.

[0145] (2) Judgment of recovery conditions: When the circuit breaker condition is lifted and multiple parameters (MI(t) recovers, skin temperature decreases, abdominal tension decreases) are simultaneously met, the circuit breaker is gradually lifted.

[0146] Example of recovery conditions: MI(t) recovered from -0.1 at the time of the circuit breaker to above 0.1.

[0147] Skin temperature dropped from 45°C to below 40°C.

[0148] Abdominal tension decreased from 35 kPa to below 20 kPa.

[0149] (3) Tiered recovery process: Software circuit breaker recovery: Automatic recovery, no manual intervention required.

[0150] Hardware fuse recovery: Manual reset by medical staff is required to ensure safety.

[0151] After recovery, the carrier frequency and treatment parameters are recalculated to avoid triggering the circuit breaker again.

[0152] The tiered recovery mechanism after circuit breaker interruption ensures the continuity and safety of treatment, and improves the reliability of the system and the user experience.

[0153] Appendix 1: Standardized Table of Acupoint Parameters acupoints Electrical stimulation frequency Airbag pressure (kPa) Temperature (°C) Tian Shu 2-5Hz rarefaction and compression waves 5-10 kPa 42-45℃ Zusanli 1-3Hz low frequency pulse 3-8kPa 40-43℃ Appendix 2: BIA Scan Parameter Table frequency band Electrode spacing (cm) Measurement range (Ω) Resolution (Ω) 5kHz 3cm 50-5000 1 50kHz 3cm 50-5000 1 It should be noted that, in this document, relational terms such as "one" and "two" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, the phrase "comprising an element defined as..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0154] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A postoperative gastrointestinal motility closed-loop management system, characterized in that, include: Wireless distributed sensor nodes are attached to key acupoints on the abdomen to collect bowel sounds, slow waves of gastrointestinal electrophysiology, abdominal tension, and IMU motion posture. The edge computing control terminal runs an adaptive multimodal fusion algorithm, dynamically updates weights through Kalman filtering or variational Bayesian inference, and calculates the gastrointestinal motility index MI(t). The dual-level safety execution module includes software logic fuses, namely, stopping when the algorithm detects an abnormality, and hardware physical fuses, namely, MOSFET relays forcibly cutting off power; The cloud-based monitoring platform supports remote visual monitoring, data traceability, and OTA upgrades, and uses AES-256 hardware encryption. The system implements a four-layer closed-loop architecture encompassing perception, decision-making, execution, and monitoring. The calculation formula is: ； in, Energy for bowel sounds, For slow wave rhythm regularity, For abdominal tension, This is the IMU motion compensation term.

2. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The adaptive carrier frequency adjustment module scans the abdominal wall fat thickness using BIA and adaptively sets the carrier frequency based on the abdominal wall fat thickness. The mapping formula is as follows: ； in By amplifying the high impedance difference through natural logarithm, the penetration of the device is optimized for obese patients.

3. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The wireless distributed sensing node includes at least two sensing units, which are respectively attached to the Tianshu acupoint and the Zusanli acupoint.

4. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The circuit breaker threshold includes: A skin temperature ≥45°C triggers a hardware power-off. A higher-level circuit breaker is triggered when MI(t) remains below the threshold and the heart rate variability (HRV) is abnormal. The dynamic adjustment model for the circuit breaker threshold is as follows: ； in , The lower the impedance or the higher the trunk impedance, the earlier the fuse threshold is triggered.

5. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The cloud-based monitoring platform uses AES-256 hardware encryption, supports OTA upgrade signature verification and anti-rollback, and the encryption protocol combines HIPAA compliance with a federated learning framework. Each hospital stores patient data locally and only uploads encrypted MI(t) model increments.

6. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The system includes a tiered recovery mechanism after circuit breaker failure, which includes: After the hardware fuse fails, the system automatically enters "security monitoring mode"; Continuous monitoring of the patient's condition using low-power sensors; The circuit breaker will be gradually deactivated when multiple parameters meet the conditions simultaneously.

7. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The system includes a graded alarm system for safety fuse tripping, the system comprising: Level 1 software circuit breaker triggers an alarm at the nurse station. A secondary hardware fuse failure triggers an audible and visual alarm. The system establishes a correspondence between alarm levels and circuit breaker levels, and supports remote visual monitoring, data traceability, and OTA upgrades.

8. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The system includes an adaptive electrical stimulation waveform unit driven by bowel phonation spectral characteristics, which comprises: By analyzing bowel sounds using FFT, metallic frequency and rhythm disorder bands were extracted. Establish a mapping table between spectral characteristics and electrical stimulation parameters; Edge computing terminals analyze the spectrum in real time, dynamically adjust the waveform, and record therapeutic feedback.

9. The postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The system includes a closed-loop control unit for joint decision-making based on traditional Chinese medicine syndrome differentiation and physiological parameters. This unit includes: Combining data from the four diagnostic methods of traditional Chinese medicine with quantitative indicators of postoperative gastrointestinal motility; Construct a joint scoring model of "syndrome type-physiological parameters"; The edge computing terminal updates the certificate score every hour and adjusts the electrical, mechanical and thermal parameters accordingly.

10. A postoperative gastrointestinal motility closed-loop management system according to claim 1, characterized in that: The system also includes a flexible piezoelectric sensor array integration unit, which comprises: Deploy a PVDF nanofiber piezoelectric sensor array in the abdominal sensing node; Simultaneous acquisition of acoustic vibration and abdominal tension; Real-time calibration is performed using an edge computing terminal to offset changes in electrode contact impedance.

Images