Intelligent stomach tube positioning system and positioning method thereof

Abstract

The invention relates to an intelligent stomach tube positioning system and a positioning method thereof, and belongs to the technical field of medical stomach tubes, the system comprises a stomach tube body, the outer side of the stomach tube body is provided with an FBG sensor, and the FBG sensor is used for collecting deformation signals of the inner wall of the esophagus in real time; the data processing unit is in signal connection with the FBG sensor, a deformation waveform classification model is arranged in the data processing unit, three physiological stenosis sections of the esophagus can be identified by analyzing deformation signals, and the three physiological stenosis sections are the annular pharyngeal stenosis section, the bronchial intersection stenosis section and the diaphragm stenosis section respectively; the verification module comprises a pH sensor and an air pressure sensor which are arranged on the stomach tube body and is used for verifying the in-place state of the stomach tube body in a multi-mode manner; and the interaction module is in signal connection with the data processing unit and the verification module and is used for displaying the passing state of the physiological narrow section and triggering a multi-level alarm. According to the stomach tube, the position of the stomach tube body can be recognized, pain brought to a patient due to repeated insertion of the stomach tube in a traditional blind insertion mode is reduced, and clinical operation safety is improved.

Description

Technical Field

[0001] This application relates to the field of medical gastric tubes, and in particular to an intelligent gastric tube positioning system and its positioning method. Background Art

[0002] Currently, in clinical practice, patients who cannot take food orally or need gastric lavage all need to have a gastric tube inserted. Clear water, food, or cleaning agents are injected through the gastric tube, or the gastric juice after washing is extracted through the gastric tube. When in use, the gastric tube enters the human body through the nasal cavity and then through the esophagus, so that one end of the gastric tube is inserted into the patient's stomach, and then clear water, food, or cleaning fluid is injected into the stomach through the gastric tube, or the gastric juice is extracted.

[0003] A related technology discloses a gastric tube, which increases its toughness by placing a guide wire in the lumen of the gastric tube body and does not affect its softness, so as to effectively improve the success rate of the first insertion during the insertion operation. By setting scales on the outer wall of the gastric tube body, medical staff can accurately operate the insertion length of the gastric tube, thereby effectively reducing the pain of the patient.

[0004] In view of the above, although the insertion depth of the gastric tube can be visually seen through the scale, the specific position of the insertion end of the gastric tube (in the stomach, oral cavity, or esophagus) cannot be determined. For example, for adults of 1.5 meters and 1.9 meters, when the gastric tube is inserted to the same depth, the specific positions of the insertion ends of the gastric tubes are not the same, and it is difficult to judge whether the insertion end of the gastric tube is placed in the gastric cavity, which is not conducive to the development of medical care work. Summary of the Invention

[0005] In order to facilitate the determination of the specific position of the insertion end of the gastric tube, so as to avoid the pain brought to the patient and the workload of medical staff due to repeated insertion of the gastric tube, this application provides an intelligent gastric tube positioning system and its positioning method.

[0006] In a first aspect, this application provides an intelligent gastric tube positioning system, adopting the following technical solution:

[0007] An intelligent gastric tube positioning system includes:

[0008] A gastric tube body, on which an FBG sensor is provided, and the FBG sensor is used to collect the deformation signals of the inner wall of the esophagus in real time;

[0009] A data processing unit, which is signal-connected to the FBG sensor and has a deformation waveform classification model built in, and can identify three physiological stenosis segments of the esophagus by analyzing the deformation signals. The three physiological stenosis segments are the pharyngeal constriction segment, the bronchial crossing constriction segment, and the diaphragmatic constriction segment;

[0010] The verification module, including a pH sensor and a pressure sensor disposed on the gastric tube body, is used for multi-modal verification of the in-place state of the gastric tube body;

[0011] The interaction module is respectively connected to the data processing unit and the verification module in a signal manner, and is used for displaying the passing state of the physiological stenosis segment and triggering multi-level alarms.

[0012] By adopting the above technical solution, the present application based on the FBG sensor collects the esophageal deformation signal in real time, combines the deformation waveform classification model in the data processing unit to identify the physiological stenosis segment of the esophagus, and uses the multi-modal verification module (pH and air pressure) for cross-verification, which can realize the accurate dynamic judgment of the position of the gastric tube body, is beneficial to reducing the pain brought to the patient by repeated insertion of the gastric tube in traditional blind insertion and the workload of medical staff, and improves the safety of clinical operations.

[0013] Optionally, a plurality of the FBG sensors are provided, and the plurality of FBG sensors are circumferentially distributed along the gastric tube body. The detection coverage angle of a single FBG sensor is 100°

[0014] —150°, and there is a detection overlapping area of 20°—40° between adjacent sensors.

[0015] By adopting the above technical solution, through the layout design of the circumferentially distributed and overlapping FBG sensors, the three-dimensional deformation of the esophagus can be monitored without dead angles, which is beneficial to avoiding signal omission caused by too large sensor spacing, adapting to the anatomical structure differences of patients of different ages (adults / newborns), and ensuring the integrity of the identification of the physiological stenosis segment.

[0016] Optionally, the data processing unit includes a preprocessing subunit and a classification subunit. The preprocessing subunit uses an adaptive wavelet basis function to denoise the deformation signal, and the classification subunit outputs the classification result of the physiological stenosis segment in real time by deploying a deformation waveform classification model.

[0017] By adopting the above technical solution, the present application uses an adaptive wavelet basis function for denoising in the data processing unit, and the classification subunit outputs the classification result of the physiological stenosis segment in real time, which can significantly improve the signal-to-noise ratio, is beneficial to reducing positioning errors caused by signal interference, and further enhances the system robustness.

[0018] Optionally, the interaction module includes a primary alarm unit and a secondary alarm unit. Among them, when the pressure sensor detects that the gastric tube body deviates from the esophageal path, the primary alarm unit is triggered; when the pH sensor detects that the pH value is in the critical range of 4.0—4.5, the secondary alarm unit is triggered.

[0019] By adopting the above technical solution, a hierarchical alarm mechanism is adopted, which can differentially prompt medical staff according to the risk level, is beneficial to shorten the response time in emergency misinsertion scenarios (such as sudden air pressure drop), and avoid excessive alarm interference at the critical pH value, thereby optimizing the human-machine interaction efficiency.

[0020] Optionally, the verification module further includes a time synchronization unit, which is respectively connected to the FBG sensor, the pH sensor and the air pressure sensor for hardware-level time alignment of the signals of the FBG sensor, the pH sensor and the air pressure sensor.

[0021] By adopting the above technical solution, by aligning the multi-modal signal timings through the time synchronization unit, the logical conflicts caused by the delay between sensors (such as the signal detected by the pH sensor lagging behind the signal detected by the FBG sensor) can be eliminated, which is beneficial to solving the misjudgment problem caused by the delay of multi-sensor signals, and thus beneficial to maintaining the system reliability in the high-speed intubation scenario.

[0022] In a second aspect, the present application also provides a positioning method for an intelligent gastric tube positioning system, adopting the following technical solution:

[0023] A positioning method for an intelligent gastric tube positioning system includes the following steps:

[0024] S1. Collect the deformation signal of the esophageal inner wall through the FBG sensor and preprocess the deformation signal;

[0025] S2. Identify the physiological stenosis segments of the esophagus and the passing order through the deformation waveform classification model;

[0026] S3. If the passing order in S2 is successively the pharyngeal stenosis segment, the bronchial crossing stenosis segment and the diaphragmatic stenosis segment, it is determined that the gastric tube body is close to the stomach;

[0027] S4. Start pH verification. If the pH value is detected to be lower than 4.0 twice in a row, it is confirmed that the gastric tube body has reached the stomach.

[0028] By adopting the above technical solution, through the collaborative logic of sequential recognition of physiological stenosis segments and double pH verification, it is beneficial to improve the accuracy of judging the position of the gastric tube, reduce misjudgment caused by the failure of a single sensor, and thus reduce the probability of secondary intubation.

[0029] Optionally, the training method of the deformation waveform classification model in S2 includes: adding an anatomical order penalty term to the loss function, and if the output order is incorrect, the loss value automatically increases; adding simulated clinical noise to neonatal data to enhance generalization.

[0030] By adopting the above technical solutions, in the model training of the present application, an anatomical order penalty term and respiratory noise data augmentation are introduced, which can force the classification results to conform to the physiological path logic, improve the generalization of neonatal stenosis segment detection, help solve the overfitting problem under small-sample data, and further help reduce the misjudgment rate of neonatal gastric tube placement judgment.

[0031] Optionally, if the pH verification fails in S4, the following steps are executed:

[0032] S41. Detect the gastric peristalsis signal: perform band-pass filtering on the deformation signal at 0.05 - 0.1 Hz and calculate the power spectral density;

[0033] S42. If the peak value of the power spectral density is greater than 6 dB, it is determined that the gastric tube body has reached the stomach.

[0034] By adopting the above technical solutions, when the pH verification fails in the present application, the determination result is covered by detecting the gastric peristalsis signal, which can avoid false detection caused by abnormal gastric acid secretion, help maintain the positioning reliability in complex pathological scenarios (such as gastric acid reflux), and further reduce unnecessary adjustment operations of the gastric tube body position.

[0035] Optionally, the physiological stenosis segment dynamic calibration mechanism is included in S2: every time a physiological stenosis segment is recognized, the frequency domain feature threshold for subsequent detection is adjusted in real time; the calibration parameters are dynamically selected according to the patient's age.

[0036] By adopting the above technical solutions, the present application adopts the physiological stenosis segment dynamic calibration mechanism, which can adaptively optimize the detection logic according to individual anatomical differences, help solve the recognition problem of neonates, and further improve the accuracy of recognizing physiological stenosis segments for patients of different ages.

[0037] Optionally, the abnormal path self-correction mechanism is included in S3: if no physiological stenosis segment is recognized and the air pressure drops suddenly by more than 50%, it is determined that the gastric tube body has entered the airway by mistake; a path prediction model is constructed based on historical misinsertion data to generate a reverse propulsion path.

[0038] By adopting the above technical solutions, the present application constructs a path prediction model based on historical misinsertion data, which can reverse calculate the path where the gastric tube body has entered by mistake and generate a withdrawal guidance, with an active error correction effect, which can effectively shorten the operation time and further significantly reduce the risk of secondary tissue damage.

[0039] In summary, the present application includes the following beneficial technical effects:

[0040] 1. This application is based on real-time collection of esophageal deformation signals by FBG sensors, combines the deformation waveform classification model in the data processing unit to identify the physiological stenosis segment of the esophagus, and uses a multimodal verification module (pH and air pressure) for cross-verification, enabling accurate dynamic determination of the position of the gastric tube body. This helps reduce the pain suffered by patients and the workload of medical staff caused by repeated insertion of the gastric tube during traditional blind insertion, and improves the safety of clinical operations. Description of the Drawings

[0041] Figure 1 is a structural block diagram of an intelligent gastric tube positioning system according to an embodiment of the present application.

[0042] Figure 2 is a sectional view of the gastric tube body showing an embodiment of the present application.

[0043] Figure 3 is a physiological structure diagram of the esophagus showing an embodiment of the present application.

[0044] Figure 4 is a structural block diagram of the data processing unit showing an embodiment of the present application.

[0045] Figure 5 is a structural block diagram of the verification module showing an embodiment of the present application.

[0046] Figure 6 is a structural block diagram of the interaction module showing an embodiment of the present application.

[0047] Figure 7 is a flowchart of a positioning method for an intelligent gastric tube positioning system according to an embodiment of the present application.

[0048] Figure 8 is a comparison chart showing the power spectral density of signals from the stomach, esophagus, and intestine.

[0049] Description of Reference Numerals: 1, gastric tube body; 2, FBG sensor; 3, data processing unit; 31, preprocessing subunit; 32, classification subunit; 4, verification module; 41, pH sensor; 42, air pressure sensor; 43, time synchronization unit; 5, interaction module; 51, primary alarm unit; 52, secondary alarm unit; 6, physiological stenosis segment; 61, cricopharyngeal stenosis segment; 62, bronchial crossing stenosis segment; 63, diaphragmatic stenosis segment. Detailed Embodiments

[0050] The following is a more detailed description of the present application in conjunction with Figure 1 — Figure 8 to further elaborate on the present application.

[0051] Embodiment:

[0052] In a first aspect, an embodiment of the present application discloses an intelligent gastric tube positioning system.

[0053] Reference Figure 1 and Figure 2 , an intelligent gastric tube positioning system, comprising a gastric tube body 1, an FBG sensor 2, a data processing unit 3, a verification module 4 and an interaction module 5. Each module forms a closed-loop control system through physical connection and signal transmission: during the intubation process of the gastric tube body 1, the deformation signal is collected by the FBG sensor 2, and after being analyzed by the data processing module, the verification module 4 performs multimodal cross-validation, and finally the positioning result and alarm information are fed back through the interaction module 5.

[0054] Reference Figure 1 The stomach tube body 1 is made of biocompatible silicone material, and the FBG sensor 2 is fixed to the outer wall of the insertion end of the stomach tube body 1, which is used to collect the deformation signal of the inner wall of the esophagus in real time. The FBG sensor is also called a fiber Bragg grating (FiberBragg Grating) sensor, which can achieve accurate measurement of physical quantities such as temperature, strain, and pressure by selectively reflecting light of a specific wavelength (Bragg wavelength). There are multiple FBG sensors 2, and the multiple FBG sensors 2 are distributed along the circumference of the stomach tube body 1. The detection coverage angle of a single sensor is 100°-150°, and there is a detection overlap area of 20°-40° between adjacent sensors. In this way, through the layout design of the FBG sensors 2 with circumferential distribution and overlapping coverage, it is possible to achieve blind spot monitoring of the three-dimensional deformation of the esophagus, which is conducive to avoiding signal leakage caused by excessive sensor spacing, adapting to the anatomical structure differences of patients of different ages (adults / neonates), and ensuring the integrity of the identification of the physiological stenosis segment 6.

[0055] Reference Figure 1 and Figure 3 The data processing unit 3 is connected to the FBG sensor 2 by signal, and has a built-in deformation waveform classification model, which can identify three physiological stenosis segments 6 of the esophagus by analyzing the deformation signal. The three physiological stenosis segments 6 are the cricopharyngeal stenosis segment 61, the bronchial intersection stenosis segment 62 and the diaphragmatic stenosis segment 63, which are specifically described as follows:

[0056] Reference Figure 4 The data processing unit 3 includes a preprocessing subunit 31 and a classification subunit 32. The preprocessing subunit 31 uses an adaptive wavelet basis function to denoise the deformation signal, and the classification subunit 32 outputs the classification result of the physiological stenosis segment 6 in real time by deploying a deformation waveform classification model. In this way, since the original deformation signal collected by the FBG sensor 2 is easily affected by respiratory interference, intubation friction, etc., the present application uses an adaptive wavelet basis function in the data processing unit 3 to denoise the signal, and outputs the classification result of the physiological stenosis segment 6 in real time through the classification subunit 32, which can significantly improve the signal-to-noise ratio and help reduce positioning errors caused by signal interference.

[0057] The deformation waveform classification model is a machine learning-based algorithm model specifically used to identify the deformation signal waveforms collected by fiber Bragg grating (FBG) sensors and classify the corresponding physiological structures or external action types. In the intelligent gastric tube positioning system of this application, the core function of this model is to identify the characteristic waveforms generated when the gastric tube passes through the three physiological stenosis segments 6 (the pharyngeal stenosis segment 61, the bronchial crossing stenosis segment 62, and the diaphragmatic stenosis segment 63) of the esophagus, so as to determine the real-time position of the gastric tube body 1.

[0058] Refer to Figure 5 , the verification module 4 integrates a pH sensor 41, a pressure sensor 42, and a time synchronization unit 43, and is used to verify the in-place state of the gastric tube body 1 in a multi-modal manner. Among them, the pH sensor 41 is located at the front end of the side wall of the gastric tube body 1, and the detection range is 1.0 - 9.0; the pressure sensor 42 monitors the pressure change inside the gastric tube body 1. The time synchronization unit 43 uses a hardware-level timestamp chip and is respectively signal-connected to the FBG sensor 2, the pH sensor 41, and the pressure sensor 42, and is used to synchronize and align the signals of the FBG sensor 2, the pH sensor 41, and the pressure sensor 42. The alignment error ≤ 5ms, and the signal consistency is verified through the following logic:

[0059] When the FBG sensor 2 detects the diaphragmatic stenosis segment 63, the pH sensor 41 needs to start detection within 1 second;

[0060] When the pressure sensor 42 detects a pressure sudden drop > 50%, the output signal of the FBG sensor 2 needs to show abnormal fluctuations within 0.2 seconds.

[0061] In this way, by aligning the multi-modal signal timings through the time synchronization unit 43, the logical conflicts caused by the delay between sensors (such as the signal detected by the pH sensor 41 lagging behind the signal detected by the FBG sensor 2) can be eliminated, which is beneficial to solving the misjudgment problem caused by the signal delay of multiple sensors, and further beneficial to maintaining the system reliability in the high-speed intubation scenario.

[0062] Refer to Figure 6, the interaction module 5 includes a first-level alarm unit 51, a second-level alarm unit 52, and a visualization interface 53. Among them, the first-level alarm unit 51 uses a high-frequency sound and light alarm, and the second-level alarm unit 52 uses a low-frequency buzzer for different warning distinctions. And when the air pressure sensor 42 detects a sudden change in the air pressure value and determines that the gastric tube body 1 deviates from the esophageal path, the first-level alarm unit 51 is triggered; when the pH sensor 41 detects that the pH value is in the critical range of 4.0 - 4.5, the second-level alarm unit 52 is triggered; the visualization interface 53 is a display screen, which can display the passing state of the physiological stenosis segment 6 in real time according to the classification result output by the classification subunit 32. In this way, by adopting a hierarchical alarm mechanism, it is possible to differentially prompt medical staff according to the risk level, which is beneficial to shortening the response time in emergency misinsertion scenarios (such as sudden air pressure drop), and at the same time avoiding excessive alarm interference of the critical pH value, thereby optimizing the human-computer interaction efficiency.

[0063] The implementation principle of an intelligent gastric tube positioning system according to an embodiment of the present application is as follows: Based on the FBG sensor 2, the present application collects esophageal deformation signals in real time, combines the deformation waveform classification model in the data processing unit 3 to identify the physiological stenosis segment 6 of the esophagus, and uses the multi-modal verification module 4 (pH and air pressure) for cross-verification, which can realize the accurate dynamic judgment of the position of the gastric tube body 1, is beneficial to reducing the pain caused to patients and the workload of medical staff by repeatedly inserting the gastric tube in traditional blind insertion, and improving the safety of clinical operations.

[0064] In the second aspect, the present application also discloses a positioning method for an intelligent gastric tube positioning system.

[0065] Referring to Figure 7 , a positioning method for an intelligent gastric tube positioning system includes the following steps:

[0066] S1. Collect the deformation signals of the esophageal inner wall through the FBG sensor 2 and preprocess the deformation signals;

[0067] S2. Identify the physiological stenosis segment 6 of the esophagus and the passing sequence through the deformation waveform classification model;

[0068] S3. If the passing sequence in S2 is successively the pharyngeal constriction segment 61, the bronchial crossing constriction segment 62, and the diaphragmatic constriction segment 63, it is determined that the gastric tube body 1 is close to the stomach;

[0069] S4. Start pH verification. If the pH value is continuously detected to be lower than 4.0 twice, it is confirmed that the gastric tube body 1 has reached the stomach.

[0070] This positioning method through the collaborative logic of sequential recognition of the physiological stenosis segment 6 and double pH verification is beneficial to improving the accuracy of gastric tube placement judgment, reducing misjudgment caused by the failure of a single sensor, and thus reducing the probability of secondary intubation.

[0071] Specifically, in S1, the preprocessing of the signal includes the following processing steps:

[0072] S11. Denoise through an adaptive wavelet basis function to eliminate breathing interference and intubation friction;

[0073] S12. Synchronize and align the output signals of the FBG sensor 2, pH sensor 41, and barometric pressure sensor 42 through the time synchronization unit 43 to ensure strict synchronization of multimodal data.

[0074] Specifically, in S2, the preprocessed deformation signal is subjected to short-time Fourier transform (STFT) to generate a time-frequency feature matrix, which is input into the deformation waveform classification model for classification. The deformation waveform classification model outputs the types of the physiological stenosis segments 6 (pharyngeal stenosis segment 61, bronchial crossing stenosis segment 62, and diaphragmatic stenosis segment 63) and the passing order.

[0075] The training method of the above deformation waveform classification model includes:

[0076] S21. Add an anatomical order penalty term to the loss function. If the output order is incorrect, the loss value automatically increases, specifically as follows:

[0077] S211. Loss function design: Add an order penalty term on the basis of the cross-entropy loss to obtain formula ①:

[0078] In formula ①, L CE : Standard cross-entropy loss; α: Penalty coefficient; φ: Indicator function, taking 1 when the predicted order does not conform to the physiological order (such as identifying the diaphragmatic stenosis segment 63 first and then the pharyngeal stenosis segment 61), otherwise taking 0; y t : Predicted category at the current moment; Expected category at the next moment (determined according to the anatomical order).

[0079] S212. Dynamic weight adjustment: If the penalty coefficient has two consecutive incorrect prediction orders, automatically increase the value of α to strengthen the anatomical order constraint.

[0080] S22. Add simulated clinical noises (such as breathing noises and intubation friction noises) to the neonatal data to enhance generalization.

[0081] By introducing an anatomical order penalty term and clinical noise data augmentation in the training of the deformation waveform classification model, it is possible to force the classification result to conform to the physiological path logic, improve the generalization of neonatal stenosis segment detection, facilitate solving the overfitting problem under small sample data, and further facilitate reducing the misjudgment rate of neonatal gastric tube placement judgment.

[0082] S2 also includes a dynamic calibration mechanism for physiological stenosis segments: every time a physiological stenosis segment 6 is recognized, the frequency domain feature threshold for subsequent detections is adjusted in real time; the calibration parameters are dynamically selected according to the patient's age, as follows:

[0083] S23. After the pharyngeal constriction segment 61 is recognized, the frequency domain feature threshold of the bronchial crossing constriction segment 62 is reduced by 10% (for adults) or 20% (for neonates), and the calibration parameter is calculated by formula ②:

[0084] Formula ②: T NEW = T BASE ·(1 - β1·A), where β1 = 0.1 (for adults) or 0.2 (for neonates), and A is the age weight coefficient;

[0085] In this way, when the system recognizes the pharyngeal constriction segment 61, the frequency domain feature threshold of the bronchial crossing constriction segment 62 is reduced to enhance the sensitivity to airway misentry and further strengthen the early warning of misinsertion risk.

[0086] S24. After the bronchial crossing constriction segment 62 is recognized, the frequency domain feature threshold of the diaphragmatic constriction segment 63 is increased by 15% (for adults) or 25% (for neonates), and the calibration parameter is calculated by formula ③:

[0087] Formula ③: T NEW = T BASE ·(1 + β2·A), where β2 = 0.15 (for adults) or 0.25 (for neonates), and A is the age weight coefficient.

[0088] In this way, after the bronchial crossing constriction segment 62 is detected, the frequency domain feature threshold of the diaphragmatic constriction segment 63 is increased to compensate for the influence of respiratory interference, ensuring that the positioning is triggered only when the diaphragmatic constriction segment 63 is truly reached.

[0089] Specifically, S3 includes an abnormal path self-correction mechanism:

[0090] S31. If no physiological stenosis segment 6 is recognized and the sudden drop in air pressure is greater than 50%, it is determined that the gastric tube body 1 has entered the airway by mistake;

[0091] S32. Based on historical misinsertion data, a path prediction model is constructed to generate a reverse propulsion path;

[0092] S33. Prompt the medical staff to retract the gastric tube body 1 along the reverse propulsion path, and then re-push the gastric tube body 1 forward directly until the gastric tube body 1 reaches the stomach.

[0093] In this way, based on historical misinsertion data, the present application constructs a path prediction model, which can reverse calculate the misentry path of the gastric tube body 1 and generate a retraction guide, having an active error correction effect, which can effectively shorten the operation time and further significantly reduce the risk of secondary tissue damage.

[0094] Specifically, if the pH verification fails in S4, the following steps are executed:

[0095] S41. Detect the gastric peristalsis signal: perform band-pass filtering on the collected deformation signal at 0.05 - 0.1 Hz, and calculate the power spectral density;

[0096] S42. If the peak value of the power spectral density is greater than 6 dB, it is determined that the gastric tube body 1 has reached the stomach, and the specific basis is as follows:

[0097] Physiologically, the basic electrical rhythm frequency of gastric peristalsis is between 0.05 - 0.1 Hz; in the frequency band of 0.05 - 0.1 Hz, referring to Figure 8 , the gastric signal power is significantly higher than that of the esophagus (almost no spontaneous peristalsis) and the intestine (higher frequency, 0.12 - 0.2 Hz). Therefore, setting the peak value of the power spectral density to be greater than 6 dB can exclude the vast majority of esophageal or intestinal interference signals.

[0098] In this way, when the pH verification fails, the present application detects and covers the determination result through the gastric peristalsis signal, which can avoid false detection caused by abnormal gastric acid secretion, is beneficial to maintaining the positioning reliability in complex pathological scenarios (such as gastric acid reflux), and further reduces unnecessary position adjustment operations of the gastric tube body 1.

[0099] The implementation principle of the positioning method of the intelligent gastric tube positioning system in the embodiment of the present application is as follows: this positioning method encodes the physiological structure characteristics of the esophagus into strain waveform characteristics recognizable by the positioning system, and through the sequential recognition of the physiological stenosis section 6 and the collaborative logic of double pH verification, it is beneficial to improve the accuracy of the judgment of the gastric tube body 1 "reaching the stomach", reduce misjudgment caused by the failure of a single sensor, and further reduce the probability of secondary intubation. Its core value lies in the deep integration of anatomical knowledge and biomedical signal processing technology, realizing the quantification, intelligence, and generalization of the navigation of the gastric tube body 1.

[0100] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of the present application should be covered within the protection scope of the present application.

Claims

1. An intelligent gastric tube positioning system, characterized in that, Including: A gastric tube body (1), on the outer side of which there is an FBG sensor (2), and the FBG sensor (2) is used to collect the deformation signals of the esophageal inner wall in real time; A data processing unit (3), which is signal-connected to the FBG sensor (2), and has a built-in deformation waveform classification model, and can identify three physiological stenosis segments (6) of the esophagus by analyzing the deformation signals. The three physiological stenosis segments (6) are respectively the cricopharyngeal stenosis segment (61), the bronchial crossing stenosis segment (62), and the diaphragmatic stenosis segment (63); A verification module (4), including a pH sensor (41) and a pressure sensor (42) arranged on the gastric tube body (1), for multi-modal verification of the in-place state of the gastric tube body (1); An interaction module (5), which is respectively signal-connected to the data processing unit (3) and the verification module (4), and is used to display the passing state of the physiological stenosis segment (6) and trigger multi-level alarms.

2. The intelligent gastric tube positioning system according to claim 1, wherein: There are multiple FBG sensors (2), and the multiple FBG sensors (2) are distributed along the circumferential direction of the gastric tube body (1). The detection coverage angle of a single FBG sensor (2) is 100° - 150°, and there is a detection overlapping area of 20° - 40° between adjacent sensors.

3. The intelligent gastric tube positioning system according to claim 1, characterized in that: The data processing unit (3) includes a preprocessing subunit (31) and a classification subunit (32). The preprocessing subunit (31) uses an adaptive wavelet basis function to denoise the deformation signal, and the classification subunit (32) outputs the classification result of the physiological stenosis segment (6) in real time by deploying a deformation waveform classification model.

4. An intelligent gastric tube positioning system according to claim 1, characterized in that: The interaction module (5) includes a first-level alarm unit (51) and a second-level alarm unit (52). Among them, when the pressure sensor (42) detects that the gastric tube body (1) deviates from the esophageal path, the first-level alarm unit (51) is triggered; when the pH sensor (41) detects that the pH value is in the critical range of 4.0 - 4.5, the second-level alarm unit (52) is triggered.

5. An intelligent gastric tube positioning system according to claim 1, characterized in that: The verification module (4) further includes a time synchronization unit (43), and the time synchronization unit (43) is respectively signal-connected to the FBG sensor (2), the pH sensor (41), and the pressure sensor (42), and is used to perform hardware-level time alignment on the signals of the FBG sensor (2), the pH sensor (41), and the pressure sensor (42).

6. A positioning method of the intelligent gastric tube positioning system according to any one of claims 1-5, characterized in that, Including the following steps: S1. Collect the deformation signals of the esophageal inner wall through the FBG sensor (2) and preprocess the deformation signals; S2. Identify the physiological stenosis segments (6) of the esophagus and the passing sequence through the deformation waveform classification model; S3. If the passing sequence in S2 is successively the cricopharyngeal stenosis segment (61), the bronchial crossing stenosis segment (62), and the diaphragmatic stenosis segment (63), it is determined that the gastric tube body (1) is approaching the stomach; S4. Start pH verification, and confirm that the gastric tube body (1) reaches the stomach if the pH value is detected to be lower than 4.0 twice continuously.

7. The positioning method of an intelligent gastric tube positioning system according to claim 6, characterized in that, The training method of the deformation waveform classification model described in S2 includes: adding an anatomical order penalty term to the loss function, and if the output order is incorrect, the loss value will automatically increase; adding simulated clinical noise to neonatal data to enhance generalization.

8. The positioning method of an intelligent gastric tube positioning system according to claim 6, characterized in that, If the pH verification fails in S4, the following steps are executed: S41. Detect the gastric peristalsis signal: perform a 0.05 - 0.1 Hz band-pass filter on the deformation signal and calculate the power spectral density; S42. If the peak value of the power spectral density is greater than 6 dB, it is determined that the gastric tube body (1) has reached the stomach.

9. The positioning method of an intelligent gastric tube positioning system according to claim 6, characterized in that, S2 includes a dynamic calibration mechanism for physiological stenosis segments: every time a physiological stenosis segment (6) is recognized, the frequency domain feature threshold for subsequent detection is adjusted in real time; the calibration parameters are dynamically selected according to the patient's age.

10. The positioning method of an intelligent gastric tube positioning system according to claim 6, characterized in that, S3 includes an abnormal path self-correction mechanism: if no physiological stenosis segment (6) is recognized and the sudden drop in air pressure is greater than 50%, it is determined that the gastric tube body (1) has entered the airway by mistake; a path prediction model is constructed based on historical misinsertion data to generate a reverse propulsion path.

Images