Radial artery puncture navigation system based on multi-modal sensing and digital projection

Abstract

The invention relates to the technical field of medical instruments and clinical nursing, in particular to a radial artery puncture navigation system based on multi-modal sensing and digital projection, which comprises the following steps: a multi-modal blood vessel signal acquisition module, which is used for acquiring a blood vessel pulsation signal on the wrist of a patient through a photoplethysmography sensor array, the strongest pulsation line is positioned through the pressure sensor array, and the blood vessel depth and artery and vein distinguishing information are obtained through the bioelectrical impedance sensor array. According to the invention, through multi-modal blood vessel signal acquisition, blood vessel dynamic deformation modeling, personalized digital twinning construction and dynamic updating, projection path generation and control, and projection output and real-time error compensation, real-time visual projection of a blood vessel center line and local deformation is realized; therefore, the problem that puncture positioning is not accurate due to the fact that dynamic changes of blood vessels cannot be accurately reflected due to the fact that most traditional radial artery puncture adopts a hand feeling positioning method is solved.

Description

Technical Field

[0001] This invention relates to the fields of medical devices and clinical nursing technology, and in particular to a radial artery puncture navigation system based on multimodal sensing and digital projection. Background Technology

[0002] With the increasing demands for minimally invasive procedures and higher efficiency in clinical nursing in modern medicine, vascular puncture has become a frequent and crucial step in basic clinical operations. Traditional radial artery puncture largely relies on tactile localization, which, due to its inability to accurately reflect dynamic changes in the blood vessel, leads to inaccurate puncture positioning. In recent years, technologies such as multimodal physiological signal acquisition, digital twin modeling, and visualization projection have matured in industrial, aerospace, and smart healthcare fields, providing a technological foundation for precise navigation during vascular puncture. By combining multimodal sensing technologies such as photoplethysmography, pressure sensing, and bioelectrical impedance analysis with digital twin dynamic modeling, and utilizing real-time projection for surface visualization, it is hoped that "what you see is what you puncture" can be achieved in minimally invasive puncture procedures, improving the accuracy and controllability of the operation, thereby promoting the development of clinical nursing procedures towards intelligence, quantification, and personalization. Summary of the Invention

[0003] To overcome the above shortcomings, this invention provides a radial artery puncture navigation system based on multimodal sensing and digital projection, which aims to improve the problem that traditional radial artery puncture mostly uses manual positioning, which cannot accurately reflect the dynamic changes of blood vessels, resulting in inaccurate puncture positioning.

[0004] In a first aspect, the present invention provides the following technical solution: a radial artery puncture navigation system based on multimodal sensing and digital projection, comprising the following steps:

[0005] The multimodal vascular signal acquisition module is used to acquire vascular pulsation signals on the patient's wrist through a photoplethysmography sensor array, locate the line of strongest pulsation through a pressure sensor array, and obtain vascular depth and distinguish arterial and vein information through a bioelectrical impedance sensor array.

[0006] The vascular dynamic deformation modeling module is used to filter, correct and normalize the multimodal signals, and fuse them to generate a three-dimensional continuous differentiable vascular dynamic model, which describes the spatial position and local deformation of the vascular centerline.

[0007] The personalized digital twin construction and dynamic update module is used to construct a digital twin vascular model based on the individual characteristics of the patient's vascular elasticity and skin thickness, and to generate a predicted vascular location in real time by combining the dynamic vascular model.

[0008] The projection path generation and control module is used to generate a continuous vascular centerline projection path and the optimal puncture point based on the digital twin model, and to control the projection output device to generate the projection in combination with feedforward control and closed-loop correction strategies.

[0009] The projection output and real-time error compensation module is used to project the projection path and target mark onto the patient's wrist skin, and correct the projection coordinates in real time according to the deviation signal, so as to realize the synchronous display of blood vessel projection and dynamic position.

[0010] By adopting the above technical solutions, multimodal vascular signal acquisition, dynamic deformation modeling of blood vessels, personalized digital twin construction and dynamic updating, projection path generation and control, and projection output and real-time error compensation are achieved, thereby realizing real-time visualization projection of the vascular centerline and local deformation. This improves the problem that traditional radial artery puncture mostly uses the manual positioning method, which cannot accurately reflect the dynamic changes of blood vessels, resulting in inaccurate puncture positioning.

[0011] Preferably, the multimodal vascular signal acquisition includes:

[0012] An array of photoplethysmography sensors is attached to the patient's wrist to collect optical pulsation signals from local blood vessels;

[0013] An array of pressure sensors is deployed on the surface of the wrist. The center line of the strongest pulsation is obtained through multi-point sampling, and a pressure spatial distribution map is established.

[0014] An array of bioelectrical impedance sensors is placed around the wrist to measure the impedance changes over time by applying a weak alternating current signal to distinguish between arteries and veins and to obtain information on vascular depth.

[0015] Photoelectric, pressure, and bioelectrical impedance signals are collected simultaneously to form unified multimodal vascular data.

[0016] Preferably, the vascular dynamic deformation modeling includes:

[0017] The acquired multimodal signals are filtered to remove high-frequency noise and low-frequency drift, and baseline correction and amplitude normalization are performed.

[0018] The processed signals are fused according to spatial location, and combined with PPG pulsation peak, pressure centerline and bioelectrical impedance depth information to generate a three-dimensional continuous vascular dynamic model describing the vascular centerline and local deformation.

[0019] The generated 3D vascular dynamic model is output to the digital twin construction module.

[0020] Preferably, the personalized digital twin construction and dynamic updating includes:

[0021] An initial digital twin vascular model was constructed based on the patient's vascular elasticity, skin thickness, and wrist shape.

[0022] The generated three-dimensional dynamic vascular model is mapped to the digital twin model, and the vascular centerline and local deformation are updated in real time.

[0023] We use continuous time series vascular dynamic data to generate predicted vascular location coordinates.

[0024] Preferably, the generation of predicted blood vessel locations includes:

[0025] Record the historical sequence of vascular centerline and local deformation to analyze vascular movement trends;

[0026] Generate predicted coordinates of blood vessel location at the next time point based on historical sequences and dynamic data;

[0027] The predicted coordinates are transmitted to the accurate projection path generation module.

[0028] Preferably, the projection path generation includes:

[0029] Receive the predicted blood vessel location coordinates generated by the digital twin model, generate a continuous projection path based on the centerline and local curvature, and mark the puncture point;

[0030] Increase path sampling points in areas with significant curvature changes to ensure path continuity;

[0031] Output the generated projection path coordinates to the projection control module.

[0032] Preferably, the projection path control includes:

[0033] Collect projection feedback sensor signals and calculate the deviation between the projected coordinates and the digital twin predicted coordinates;

[0034] Adjust the projection path coordinates according to the deviation to keep the projection path consistent with the dynamic position of the blood vessel;

[0035] The fusion weights are adjusted based on the stability of the signals from each modal sensor.

[0036] Preferably, the projection output and real-time error compensation include:

[0037] The projection path and puncture point coordinates drive a miniature scanning laser projector or a high-brightness LED array to project the path onto the patient's wrist skin.

[0038] The actual position of the projected light spot is measured using a projection feedback sensor, and the deviation signal is transmitted back to the control module.

[0039] The projection coordinates and scanning parameters are adjusted in real time according to the returned signal to keep the projection spot aligned with the center line of the blood vessel.

[0040] When the deviation exceeds the set range, the projection parameters are automatically adjusted to ensure continuous display.

[0041] Ideally, the collaboration between modules includes:

[0042] The modules for multimodal signal acquisition, vascular dynamic modeling, digital twin update, and projection path generation are integrated into a closed-loop operation system through data channels and control signals to achieve feedforward-closed-loop collaborative control.

[0043] The acquisition frequency, digital twin update cycle, and projection refresh rate are automatically adjusted according to the patient's physiological state and operating environment.

[0044] Maintain synchronization between the projection path and the dynamic position of the blood vessels to enable coordinated operation of all modules.

[0045] Secondly, the present invention provides the following technical solution: a radial artery puncture navigation method based on multimodal sensing and digital projection, comprising the following steps:

[0046] S1. The vascular pulsation signal is collected at the patient's wrist using a photoplethysmography sensor array, the line of strongest pulsation is located using a pressure sensor array, and the vascular depth and information distinguishing between arteries and veins are obtained using a bioelectrical impedance sensor array.

[0047] S2. The multimodal signals are filtered, baseline corrected and normalized, and then fused to generate a three-dimensional continuous differentiable vascular dynamic model, which describes the spatial position and local deformation of the vascular centerline.

[0048] S3. Construct a digital twin vascular model based on the individual characteristics of the patient's vascular elasticity and skin thickness, and combine the vascular dynamic model to generate a predicted vascular location in real time.

[0049] S4. Based on the digital twin model, generate a continuous vascular centerline projection path and the optimal puncture point, and combine feedforward control and closed-loop correction strategies to control the projection output device to generate the projection.

[0050] S5. Project the projection path and target mark onto the patient's wrist skin, and correct the projection coordinates in real time according to the deviation signal to achieve synchronous display of blood vessel projection and dynamic position.

[0051] The present invention has the following beneficial effects:

[0052] 1. In this invention, through multimodal vascular signal acquisition, dynamic deformation modeling of blood vessels, personalized digital twin construction and dynamic updating, projection path generation and control, and projection output and real-time error compensation, real-time visualization projection of the vascular centerline and local deformation is achieved, thereby improving the problem that traditional radial artery puncture mostly uses the manual positioning method, which cannot accurately reflect the dynamic changes of blood vessels, resulting in inaccurate puncture positioning.

[0053] 2. In this invention, a continuous three-dimensional vascular model is generated by modeling the dynamic deformation of blood vessels, thereby improving the problem that traditional radial artery puncture methods mostly use two-dimensional or static detection methods, which cannot reflect the dynamic changes of blood vessels and thus cause inaccurate puncture path judgment.

[0054] 3. In this invention, personalized digital twin construction and dynamic updates are used to predict the location of blood vessels based on individual patient characteristics, thereby improving the problem of mismatched puncture procedures caused by the traditional radial artery puncture method, which mostly adopts a uniform operating standard and does not take into account patient differences.

[0055] 4. In this invention, by generating the projection path and compensating for real-time errors, the projection of the blood vessel path and the dynamic position are displayed synchronously, thereby improving the problem that traditional radial artery puncture mostly uses visual estimation methods, which cannot adjust the path in real time, resulting in puncture deviation and operational risks. Attached Figure Description

[0056] Figure 1 This is a system flowchart of the radial artery puncture navigation system based on multimodal sensing and digital projection proposed in this invention.

[0057] Figure 2 This is a flowchart of the radial artery puncture navigation method based on multimodal sensing and digital projection proposed in this invention. Detailed Implementation

[0058] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Example 1:

[0060] In a first embodiment of the present invention, the present invention provides a radial artery puncture navigation system based on multimodal sensing and digital projection, such as... Figure 1 As shown, it includes the following steps:

[0061] The multimodal vascular signal acquisition module is used to acquire vascular pulsation signals on the patient's wrist through a photoplethysmography sensor array, locate the line of strongest pulsation through a pressure sensor array, and obtain vascular depth and distinguish arterial and vein information through a bioelectrical impedance sensor array.

[0062] Furthermore, multimodal vascular signal acquisition includes:

[0063] An array of photoplethysmography sensors is attached to the patient's wrist to collect optical pulsation signals from local blood vessels;

[0064] An array of pressure sensors is deployed on the surface of the wrist. The center line of the strongest pulsation is obtained through multi-point sampling, and a pressure spatial distribution map is established.

[0065] An array of bioelectrical impedance sensors is placed around the wrist to measure the impedance changes over time by applying a weak alternating current signal to distinguish between arteries and veins and to obtain information on vascular depth.

[0066] Photoelectric, pressure, and bioelectrical impedance signals are collected simultaneously to form unified multimodal vascular data.

[0067] Specifically, this module is used to acquire three key signals of blood vessels in the patient's wrist, including photoplethysmography signals, pressure signals, and bioelectrical impedance signals, to form unified multimodal vascular data, providing input data for subsequent vascular modeling, digital twins, and projection control.

[0068] A photoplethysmography (PPG) sensor array is attached to the patient's wrist. The sensors continuously measure the volume changes of the blood vessels locally, acquiring the PPG signal (PPG(t)). This signal can be expressed as: PPG(t) = I0 - I(t); where I0 is the initial light intensity received by the sensor, and I(t) is the transmitted or reflected light intensity of the blood vessel at time t. PPG(t) changes periodically with blood vessel pulsation, and a signal sequence [PPG(t1), PPG(t2), ..., PPG(t3)] is obtained through continuous sampling. n PPG signals, as inputs of vascular pulsation characteristics, provide a reference for vascular centerline localization.

[0069] A pressure sensor array is deployed on the wrist surface to perform multi-point pressure sampling, thereby obtaining the pressure centerline of the area with the strongest local pulsation. The pressure sensor array outputs pressure data P(x,y,t), where x and y represent the position coordinates of the sensor on the wrist plane, and t represents time. By calculating the local pressure gradient at each moment, the pressure centerline can be located. A pressure spatial distribution map is generated on a two-dimensional plane to aid in determining blood vessel orientation. The formula is as follows: in This is a sequence of coordinates for the points with the maximum pressure. The input of the pressure signal is data collected by the sensor, and the output is the coordinates of the pressure centerline, providing a reference for dynamic modeling of the blood vessel centerline.

[0070] An array of bioelectrical impedance sensors is placed around the wrist. By applying a weak alternating current signal, the impedance changes over time to distinguish between arteries and veins and obtain vascular depth information. The bioelectrical impedance signal is represented as: Where V(t) is the voltage after applying alternating current, I AC Z(t) represents the amplitude of the alternating current, and Z(t) varies periodically with vascular pulsation and vascular type. By analyzing the Z(t) signal in both the time and frequency domains, vascular type and depth information can be determined. The impedance signal input consists of voltage and current data acquired by the sensor, and the output is the vascular depth and arteriovenous determination results, used for multimodal data fusion.

[0071] The multimodal vascular signal acquisition module synchronously acquires the above three types of signals and aligns them according to a unified time reference to form a multimodal vascular dataset: Where D multi (t) represents the combined data collected at each time point, which serves as the input to the vascular dynamic deformation modeling module.

[0072] The entire data acquisition process includes: sensor acquisition → signal synchronization and alignment → multimodal data generation → output D multi (t). The output data includes the PPG waveform, pressure centerline coordinates, and impedance measurement results at each sampling time point, providing complete basic data for subsequent digital twin modeling, vessel location prediction, and projection control.

[0073] The vascular dynamic deformation modeling module is used to filter, correct, and normalize multimodal signals, and fuse them to generate a three-dimensional continuous differentiable vascular dynamic model, describing the spatial position and local deformation of the vascular centerline.

[0074] Furthermore, vascular dynamic deformation modeling includes:

[0075] The acquired multimodal signals are filtered to remove high-frequency noise and low-frequency drift, and baseline correction and amplitude normalization are performed.

[0076] The processed signals are fused according to spatial location, and combined with PPG pulsation peak, pressure centerline and bioelectrical impedance depth information to generate a three-dimensional continuous vascular dynamic model describing the vascular centerline and local deformation.

[0077] The generated 3D vascular dynamic model is output to the digital twin construction module.

[0078] Specifically, the module is used to perform signal preprocessing and fusion processing on multimodal vascular signals to generate a continuously differentiable three-dimensional dynamic vascular model, describing the spatial position of the vascular centerline and local deformation, and providing input data for the digital twin construction module.

[0079] The acquired photoplethysmography (PPG) signal (t), pressure signal P(x,y,t), and bioelectrical impedance signal Z(t) are filtered to remove high-frequency noise and low-frequency drift. The filtering can employ a bandpass filtering algorithm, expressed as: S f (t)=BP(S raw (t))=BP(PPG(t),P(x,y,t),Z(t)); where S raw (t) represents the acquired raw signal, S f (t) represents the filtered signal, where BP represents the bandpass filter. The filter parameters are set according to the signal sampling frequency to preserve the vascular pulsation frequency components.

[0080] The filtered signal undergoes baseline correction and amplitude normalization, as shown in the following formula:

[0081]

[0082] Where B(t) represents the baseline drift, A is the normalization coefficient, and S... n (t) represents the corrected and normalized signal sequence, used for subsequent spatial fusion.

[0083] The processed multimodal signals are fused according to their spatial location. The fusion employs a weighted spatial mapping algorithm to map the PPG pulsation peak corresponding to the vessel centerline position, pressure centerline coordinates, and impedance depth information into three-dimensional space, generating the vessel centerline and local deformation curves. The three-dimensional continuous vascular dynamic model is represented as follows: in Let X(s,t) and Y(s,t) represent the spatial position along the vessel arc length *s* at time *t*, where X(s,t) and Y(s,t) are the coordinates of the vessel centerline in the wrist plane, and Z(s,t) represents the vessel depth, determined by the impedance signal. Local deformation is obtained by calculating the displacement change along the vessel arc length direction.

[0084]

[0085] Where t0 is the initial reference time point. This represents a continuous function that describes the change of local deformation of blood vessels over time. A three-dimensional dynamic model of blood vessels is spatially continuous and differentiable, and can be used to describe the centerline position and local deformation state of blood vessels.

[0086] The input to the vascular dynamic deformation modeling module is multimodal vascular data D. multi (t); the output is a three-dimensional vascular dynamic model. and local deformation sequence The output data is passed to the digital twin construction module for subsequent digital twin model updates and projection path generation.

[0087] The data processing flow is as follows: multimodal signal input → filtering → baseline correction → amplitude normalization → spatial fusion → output of three-dimensional continuous vascular dynamic model.

[0088] This module fully discloses filtering, normalization, and multimodal fusion methods, clarifies the signal input-output relationship and mathematical representation, and provides structured, differentiable vascular dynamic data for the construction of digital twins.

[0089] The personalized digital twin construction and dynamic update module is used to construct a digital twin vascular model based on the individual characteristics of the patient's vascular elasticity and skin thickness, and combine it with the vascular dynamic model to generate predicted vascular locations in real time.

[0090] Furthermore, the construction and dynamic updating of personalized digital twins include:

[0091] An initial digital twin vascular model was constructed based on the patient's vascular elasticity, skin thickness, and wrist shape.

[0092] The generated three-dimensional dynamic vascular model is mapped to the digital twin model, and the vascular centerline and local deformation are updated in real time.

[0093] We use continuous time series vascular dynamic data to generate predicted vascular location coordinates.

[0094] Generating predicted vessel locations includes:

[0095] Record the historical sequence of vascular centerline and local deformation to analyze vascular movement trends;

[0096] Generate predicted coordinates of blood vessel location at the next time point based on historical sequences and dynamic data;

[0097] The predicted coordinates are transmitted to the accurate projection path generation module.

[0098] Specifically, this module is used to construct a digital twin vascular model based on the individual characteristics of the patient, and to update it in real time in combination with the vascular dynamic model to generate predicted vascular locations, providing continuous coordinate data for accurate projection paths.

[0099] An initial digital twin vascular model is constructed based on the patient's vascular elasticity, skin thickness, and wrist geometry. The digital twin model uses a continuous three-dimensional curve to represent the vascular centerline and incorporates vascular radius and local deformation constraints to describe the vascular spatial morphology. The initial digital twin vascular model can be represented as follows:

[0100]

[0101] in X represents the spatial position along the arc length s of the blood vessel at the initial time t0. DS (s), Y DS (s) represents the coordinates of the blood vessel in the wrist plane, Z... DS (s) represents the vascular depth, determined by vascular elasticity and skin thickness parameters. Radius and local deformation information are continuously interpolated within the model. Input data includes the patient's vascular elasticity measurement value E, skin thickness d, and wrist shape parameter W. The output is the initial digital twin model.

[0102] Model of dynamic deformation of blood vessels The data is mapped to a digital twin model, and the vessel centerline and local deformations are updated in real time. The mapping process is as follows: in This represents the local displacement of the blood vessel dynamic model over time. This is a digital twin model that is updated in real time. The module maps continuously acquired multimodal vascular data to achieve continuous updates of the time series. The input is a vascular dynamic model. The output is the updated digital twin vascular model. It is then passed to the predicted blood vessel location generation module.

[0103] Predicting vessel location formation involves analyzing the vessel centerline and local deformation history sequences. Calculate the vessel location coordinates at the next time point. The prediction formula is expressed as: in To predict the coordinates of blood vessel locations, f(·) is a prediction function based on historical local deformation sequences, used for continuous time series analysis. The input is the updated historical sequence of the digital twin model, and the output is the predicted coordinates for the next time point.

[0104] The module's data processing flow is as follows: patient individual feature input → initial digital twin model generation → multimodal vascular dynamic model mapping and updating → historical sequence recording → prediction of vascular location at the next time point → output of predicted coordinates. This flow enables continuous modeling and prediction of the vascular centerline and local deformation, providing a structured data foundation for subsequent projection.

[0105] The projection path generation and control module is used to generate continuous vascular centerline projection paths and optimal puncture points based on digital twin models, and to control the projection output device to generate projections by combining feedforward control and closed-loop correction strategies.

[0106] Furthermore, the projection path generation includes:

[0107] Receive the predicted blood vessel location coordinates generated by the digital twin model, generate a continuous projection path based on the centerline and local curvature, and mark the puncture point;

[0108] Increase path sampling points in areas with significant curvature changes to ensure path continuity;

[0109] Output the generated projection path coordinates to the projection control module.

[0110] Projection path control includes:

[0111] Collect projection feedback sensor signals and calculate the deviation between the projected coordinates and the digital twin predicted coordinates;

[0112] Adjust the projection path coordinates according to the deviation to keep the projection path consistent with the dynamic position of the blood vessel;

[0113] The fusion weights are adjusted based on the stability of the signals from each modal sensor.

[0114] Specifically, this module is used to generate a continuous projection path of the blood vessel centerline and the optimal puncture point based on a digital twin blood vessel model, and drives the projection device output through feedforward control and closed-loop correction strategies to achieve consistency between the projection path and the dynamic position of the blood vessel.

[0115] Receive the predicted blood vessel location coordinate sequence from the digital twin construction and dynamic update module. A continuous projection path is generated based on the location of the vessel's centerline and its local curvature. The continuous path can be represented as:

[0116]

[0117] in Let S be the spatial coordinates of the projection path, where S is the arc length of the blood vessel, and X is the arc length of the blood vessel. L (s), Y L (s) represents the plane projection coordinates, Z L (s) is used for projection depth reference. The path sampling interval Δs is calculated based on the vessel curvature κ(s):

[0118]

[0119] Where Δs base The baseline sampling interval is α, the curvature weighting coefficient is α, and κ(s) is the curvature function. Path continuity and projection smoothness are ensured by increasing the sampling points in areas with high curvature. The optimal puncture point is selected based on the location of the maximum local deformation along the centerline of the digital twin model and marked as the projection target point on the path. The output data is a sequence of path coordinates. and puncture point coordinates Transmitted to the projection control module.

[0120] The projection control section obtains the actual projection point coordinates by collecting signals from the projection feedback sensor. Calculate projection deviation: in To prevent coordinate deviation between the predicted path and the actual projection, the projection control module adjusts the projection device output based on the deviation. It directly converts the digital twin predicted path into projected coordinates using a feedforward control algorithm, while simultaneously correcting the deviation using closed-loop feedback to ensure the projected path aligns with the vessel location. The feedforward control input is the digital twin predicted coordinates. Closed-loop control input is deviation The output is the corrected projected coordinate sequence.

[0121] The data fusion weights w are dynamically adjusted based on the stability of the signals from each modal sensor. i The fusion formula is:

[0122]

[0123] Where i represents the three sensor modes (PPG, pressure, impedance), w represents the projection path coordinates corresponding to each mode. i Let w be the weights, satisfying ∑w i =1. By dynamically adjusting the weights, the adaptability of path generation to signal fluctuations is improved.

[0124] The entire module's data processing flow is as follows: receiving digital twin predicted coordinates → generating projection path → curvature densification sampling → marking the optimal puncture point → outputting path coordinates to projection control → acquiring feedback projection signals → calculating deviation → closed-loop correction → dynamic modal weight fusion → outputting the final projection coordinate sequence. The output data consists of continuously differentiable projection path coordinates and puncture point coordinates, providing direct input for the wristband projection device drive.

[0125] The projection output and real-time error compensation module is used to project the projection path and target mark onto the patient's wrist skin, and correct the projection coordinates in real time according to the deviation signal, so as to realize the synchronous display of blood vessel projection and dynamic position.

[0126] Furthermore, the projection output and real-time error compensation include:

[0127] The projection path and puncture point coordinates drive a miniature scanning laser projector or a high-brightness LED array to project the path onto the patient's wrist skin.

[0128] The actual position of the projected light spot is measured using a projection feedback sensor, and the deviation signal is transmitted back to the control module.

[0129] The projection coordinates and scanning parameters are adjusted in real time according to the returned signal to keep the projection spot aligned with the center line of the blood vessel.

[0130] When the deviation exceeds the set range, the projection parameters are automatically adjusted to ensure continuous display.

[0131] Specifically, this module is used to project the projection path and target coordinates onto the skin surface of the patient's wrist, and to keep the projection spot synchronized with the dynamic position of the blood vessels through real-time error compensation.

[0132] Receive projection path and puncture point coordinates and As the projection input, the projection output is achieved through a miniature scanning laser projector or a high-brightness LED array. The projection coordinates are converted into scanning control signals, which are represented as follows: in For the scan control signal, f proj f(·) is the projection coordinate transformation function, which includes projector geometric correction and scanning angle mapping. The input is the corrected projection path coordinates, and the output is the electronic control command to drive the projector.

[0133] The projection feedback section measures the actual position of the projected light spot using a sensor array. Calculate the projection error: in This indicates the deviation between the predicted path and the actual projection, with positive and negative directions representing the spatial error of deviating from the vessel centerline. Error data is transmitted back to the projection control module in real time for closed-loop correction.

[0134] Closed-loop correction achieves projection coordinate correction by adjusting the scan control signal:

[0135]

[0136] Where K is the closed-loop gain coefficient. This is the corrected projection control signal. This signal is used to drive the projection device to rescan, aligning the projected light spot with the dynamic center line of the blood vessel.

[0137] when deviation When the preset threshold is exceeded, the system automatically adjusts the projection scanning parameters, including scanning speed, step size, and beam intensity, to ensure continuous projection display and synchronization with the blood vessel position.

[0138]

[0139] Where Δs is the scan step size, v scan For scanning speed, I beam The projected light intensity is represented by adjust(·), which is an automatic adjustment function that achieves the continuity and positional accuracy of the light spot through real-time calculation.

[0140] The data processing flow is as follows: receiving corrected projection coordinates → generating scanning control signals through coordinate mapping → driving the projection device → acquiring feedback projection spot positions → calculating deviations → correcting projection coordinates in a closed loop → outputting corrected projection signals to the scanning device → automatically adjusting scanning parameters when the deviation exceeds limits → achieving synchronous projection of the blood vessel centerline. Output data includes continuous projection path coordinates and target center marker coordinates, used for real-time driving of the projection device.

[0141] The collaboration between modules includes:

[0142] The modules for multimodal signal acquisition, vascular dynamic modeling, digital twin update, and projection path generation are integrated into a closed-loop operation system through data channels and control signals to achieve feedforward-closed-loop collaborative control.

[0143] The acquisition frequency, digital twin update cycle, and projection refresh rate are automatically adjusted according to the patient's physiological state and operating environment.

[0144] Maintain synchronization between the projection path and the dynamic position of the blood vessels to enable coordinated operation of all modules.

[0145] Specifically, the system's collaborative operation includes multimodal vascular signal acquisition, dynamic vascular deformation modeling, personalized digital twin construction and dynamic updating, projection path generation and control, and projection output and real-time error compensation modules. Through data channels and control signals, a closed-loop operation system is formed to achieve feedforward-closed-loop collaborative control.

[0146] PPG signals acquired by the multimodal vascular signal acquisition module pressure signal and bioelectrical impedance signals The data is transmitted to the vascular dynamic deformation modeling module via a synchronous data bus. The modeling module performs filtering, baseline correction, and normalization on the signal to generate a three-dimensional continuous microvascular dynamic model. Describe the spatial location and local deformation of the blood vessel centerline.

[0147] Digital Twin Construction and Dynamic Update Module Receives and patient individual characteristic parameter set Generate personalized digital twin vascular models And based on continuous time series, the predicted vessel location coordinates are calculated. The coordinates are output to the projection path generation and control module, which generates the projection path based on the vessel centerline and local curvature. and puncture point coordinates The projection device is driven by feedforward control, and the deviation is controlled by closed-loop feedback. Perform path correction.

[0148] The system's coordinated control formula is expressed as: in f represents the final projection path coordinates. closed For the feedforward-closed-loop joint control function, w i The dynamic weights of each modal sensor satisfy ∑w i =1.

[0149] Based on the patient's physiological state and operating environment parameters The system adjusts the operating frequency of each module through control signals:

[0150]

[0151] Where f sample f is the sampling frequency. update For the digital twin update cycle, f proj The projection refresh rate is given by g1, g2, and g3, which are adaptive adjustment functions. The inputs are physiological and environmental parameters, and the output is the module operating frequency.

[0152] The entire module's collaborative data flow is as follows: multimodal signal acquisition → synchronous signal transmission → dynamic deformation modeling → digital twin update → predicted vessel location → projection path generation → projection control closed-loop correction → projection output → projection feedback transmission → adjustment of acquisition frequency and update cycle → feedforward-closed-loop collaborative loop. The output data includes synchronously acquired multimodal vessel data, a 3D dynamic vessel model, digital twin vessel coordinates, projection path, and puncture point coordinates, forming a complete closed-loop collaborative operation system.

[0153] Example 2:

[0154] In clinical nursing settings, healthcare professionals face challenges when performing radial artery puncture on patients, including difficulty in accurately determining vessel location, frequent dynamic changes in blood vessels, the small size or weak blood flow of vessels in the wrist, and significant individual patient differences. Traditional "blind puncture" methods relying on touch or single-modal detection struggle to reflect the vessel centerline and local deformation in real time, leading to inaccurate puncture positioning. Furthermore, single sensors are susceptible to interference and cannot simultaneously acquire information on vessel pulsation, pressure distribution, and vessel depth, resulting in deviations between the projected path and the actual vessel location. In addition, the lack of a coordinated control mechanism for multimodal signal acquisition, vessel modeling, digital twin updates, projection path generation, and projection output modules makes it impossible to adaptively adjust the acquisition frequency, digital twin update cycle, and projection refresh rate, posing a technical challenge to synchronizing the projection with the dynamic position of the vessel. To address these issues, this invention provides a radial artery puncture navigation method based on multimodal sensing and digital projection, the structure of which is as follows: Figure 2 As shown. The specific implementation process of this method is as follows:

[0155] Specifically, in the patient's wrist, an optical pulsation signal of local blood vessels is acquired using a photoplethysmography (PPG) sensor array. A pressure sensor array is deployed at multiple sampling points on the wrist surface to determine the line of strongest pulsation. A bioelectrical impedance analysis (BIA) sensor array applies a weak alternating current signal to measure the impedance changes over time, distinguishing between arteries and veins and acquiring vascular depth information. These three types of signals are acquired simultaneously to form unified multimodal vascular data. This step ensures that vascular pulsation information, pressure distribution, and vascular depth are acquired simultaneously, providing foundational data for subsequent data fusion and modeling.

[0156] The acquired multimodal signals were filtered to remove high-frequency noise and low-frequency drift, followed by baseline correction and amplitude normalization. The processed signals were then fused according to spatial location, and combined with PPG pulsatility peak value, pressure centerline, and bioelectrical impedance depth information to generate a three-dimensional continuous vascular dynamic model describing the vascular centerline and local deformation. This model can characterize the spatial location and dynamic changes of blood vessels, providing continuously differentiable vascular morphological information for the construction of digital twins.

[0157] An initial digital twin vascular model is constructed based on the patient's individual characteristics, including vascular elasticity, skin thickness, and wrist shape. The dynamic vascular model is then mapped to the digital twin model, and the vascular centerline and local deformation are updated in real time. Predicted vascular position coordinates are generated using continuous time-series data. This step achieves personalized vascular modeling and dynamic prediction, ensuring that the predicted coordinates reflect the patient's specific physiological characteristics and vascular movement trends, providing reliable input for accurate projection.

[0158] Based on the generated digital twin model, a continuous projected path of the vessel centerline is generated and the optimal puncture point is marked. For areas with significant curvature changes, additional path sampling points are added to ensure path continuity. Simultaneously, feedforward control and closed-loop correction strategies are combined to control the projection output device, adjusting the projection path coordinates in real time to ensure the generated projection path remains consistent with the vessel position predicted by the digital twin. This step transforms the digital model into directly projectable path coordinates, achieving a logical connection between projection planning and control closed-loop.

[0159] The generated projection path and puncture point coordinates drive a miniature scanning laser projector or a high-brightness LED array to project onto the patient's wrist skin. A projection feedback sensor measures the actual projection position and sends a deviation signal back to the control module. Based on the deviation signal, the projection coordinates and scanning parameters are adjusted in real time to keep the projection spot aligned with the vessel centerline. When the deviation exceeds a set range, the projection parameters are automatically corrected to maintain continuous display. This step achieves synchronous display of vessel projection and dynamic vessel position, providing an intuitive reference for medical operations and enabling closed-loop control of the projection output to ensure spatial accuracy and real-time updates of the path.

[0160] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A radial artery puncture navigation system based on multimodal sensing and digital projection, characterized in that, Includes the following steps: The multimodal vascular signal acquisition module is used to acquire vascular pulsation signals on the patient's wrist through a photoplethysmography sensor array, locate the line of strongest pulsation through a pressure sensor array, and obtain vascular depth and distinguish arterial and vein information through a bioelectrical impedance sensor array. The vascular dynamic deformation modeling module is used to filter, correct and normalize the multimodal signals, and fuse them to generate a three-dimensional continuous differentiable vascular dynamic model, which describes the spatial position and local deformation of the vascular centerline. The personalized digital twin construction and dynamic update module is used to construct a digital twin vascular model based on the individual characteristics of the patient's vascular elasticity and skin thickness, and to generate a predicted vascular location in real time by combining the dynamic vascular model. The projection path generation and control module is used to generate a continuous vascular centerline projection path and the optimal puncture point based on the digital twin model, and to control the projection output device to generate the projection in combination with feedforward control and closed-loop correction strategies. The projection output and real-time error compensation module is used to project the projection path and target mark onto the patient's wrist skin, and correct the projection coordinates in real time according to the deviation signal, so as to realize the synchronous display of blood vessel projection and dynamic position.

2. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 1, characterized in that, The multimodal vascular signal acquisition includes: An array of photoplethysmography sensors is attached to the patient's wrist to collect optical pulsation signals from local blood vessels; An array of pressure sensors is deployed on the surface of the wrist. The center line of the strongest pulsation is obtained through multi-point sampling, and a spatial distribution map of the pressure is established. An array of bioelectrical impedance sensors is placed around the wrist to measure the impedance changes over time by applying a weak alternating current signal to distinguish between arteries and veins and to obtain information on vascular depth. Photoelectric, pressure, and bioelectrical impedance signals are collected simultaneously to form unified multimodal vascular data.

3. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 1, characterized in that, The dynamic deformation modeling of blood vessels includes: The acquired multimodal signals are filtered to remove high-frequency noise and low-frequency drift, and baseline correction and amplitude normalization are performed. The processed signals are fused according to spatial location, and combined with PPG pulsation peak, pressure centerline and bioelectrical impedance depth information to generate a three-dimensional continuous vascular dynamic model describing the vascular centerline and local deformation. The generated 3D dynamic vascular model is output to the digital twin construction module.

4. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 1, characterized in that, The personalized digital twin construction and dynamic updating include: An initial digital twin vascular model was constructed based on the patient's vascular elasticity, skin thickness, and wrist shape. The generated three-dimensional dynamic vascular model is mapped to the digital twin model, and the vascular centerline and local deformation are updated in real time. We use continuous time series vascular dynamic data to generate predicted vascular location coordinates.

5. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 4, characterized in that, The generation of predicted blood vessel locations includes: Record the historical sequence of vascular centerline and local deformation to analyze vascular movement trends; Generate predicted coordinates of blood vessel location at the next time point based on historical sequences and dynamic data; The predicted coordinates are transmitted to the accurate projection path generation module.

6. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 1, characterized in that, The projection path generation includes: Receive the predicted blood vessel location coordinates generated by the digital twin model, generate a continuous projection path based on the centerline and local curvature, and mark the puncture point; Increase path sampling points in areas with significant curvature changes to ensure path continuity; Output the generated projection path coordinates to the projection control module.

7. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 6, characterized in that, The projection path control includes: Collect projection feedback sensor signals and calculate the deviation between the projected coordinates and the digital twin predicted coordinates; Adjust the projection path coordinates according to the deviation to keep the projection path consistent with the dynamic position of the blood vessel; The fusion weights are adjusted based on the stability of the signals from each modal sensor.

8. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 1, characterized in that, The projection output and real-time error compensation include: The projection path and puncture point coordinates drive a miniature scanning laser projector or a high-brightness LED array to project the path onto the patient's wrist skin. The actual position of the projected light spot is measured using a projection feedback sensor, and the deviation signal is transmitted back to the control module. The projection coordinates and scanning parameters are adjusted in real time according to the returned signal to keep the projection spot aligned with the center line of the blood vessel. When the deviation exceeds the set range, the projection parameters are automatically adjusted to ensure continuous display.

9. The radial artery puncture navigation system based on multimodal sensing and digital projection according to claim 1, characterized in that, The collaboration between modules includes: The modules for multimodal signal acquisition, vascular dynamic modeling, digital twin update, and projection path generation are integrated into a closed-loop operation system through data channels and control signals to achieve feedforward-closed-loop collaborative control. The acquisition frequency, digital twin update cycle, and projection refresh rate are automatically adjusted according to the patient's physiological state and operating environment. Maintain synchronization between the projection path and the dynamic position of the blood vessels to enable coordinated operation of all modules.

10. A radial artery puncture navigation method based on multimodal sensing and digital projection, characterized in that, The radial artery puncture navigation system based on multimodal sensing and digital projection as described in any one of claims 1-9 includes the following steps: S1. The vascular pulsation signal is collected at the patient's wrist using a photoplethysmography sensor array, the line of strongest pulsation is located using a pressure sensor array, and the vascular depth and information distinguishing between arteries and veins are obtained using a bioelectrical impedance sensor array. S2. The multimodal signals are filtered, baseline corrected and normalized, and then fused to generate a three-dimensional continuous differentiable vascular dynamic model, which describes the spatial position and local deformation of the vascular centerline. S3. Construct a digital twin vascular model based on the individual characteristics of the patient's vascular elasticity and skin thickness, and combine the vascular dynamic model to generate a predicted vascular location in real time. S4. Based on the digital twin model, generate a continuous vascular centerline projection path and the optimal puncture point, and combine feedforward control and closed-loop correction strategies to control the projection output device to generate the projection. S5. Project the projection path and target mark onto the patient's wrist skin, and correct the projection coordinates in real time according to the deviation signal to achieve synchronous display of blood vessel projection and dynamic position.

Images