Hemodialysis parameter adjusting method and millimeter wave radar

Abstract

The embodiment of the invention provides a hemodialysis parameter adjusting method and millimeter wave radar, and the method comprises the steps: measuring the vibration conditions of a blood vessel and a heart of a target object through millimeter waves, and obtaining a first signal generated by skin vibration caused by the blood vessel in an echo signal, the first signal is generated by skin vibration caused by the heart in the echo signal, the second signal is generated by skin vibration caused by the heart in the echo signal, and the first signal and the second signal pass through a signal processing model to obtain physiological state information of the target object at the current moment. According to the physiological state information at the current moment, the historical dialysis parameters of the target object and the corresponding historical physiological state information, through a dialysis parameter prediction model, the predicted dialysis parameters and / or the predicted dialysis parameter variation at the next moment are / is obtained. Under the non-contact condition, the dialysis parameters of the dialysis equipment of the target object can be timely, accurately and continuously adjusted for a long time, the method is more convenient and comfortable, the discomfort of the target object is reduced, and the method is suitable for various scenes.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method for adjusting hemodialysis parameters and millimeter-wave radar. Background Technology

[0002] Hemodialysis is a renal replacement therapy for end-stage renal disease and acute and chronic renal failure. It removes metabolic waste (urea, creatinine, etc.) and excess water from the patient's blood through diffusion, convection and ultrafiltration of semipermeable membrane media, while maintaining the body's electrolyte and acid-base balance.

[0003] A single hemodialysis session typically takes 2 to 6 hours. During this time, doctors usually need to adjust the hemodialysis parameters periodically based on the patient's current condition.

[0004] However, such adjustments usually require doctors to make them based on experience, and sometimes they cannot be made in a timely manner. Summary of the Invention

[0005] This application provides a method for adjusting hemodialysis parameters and a millimeter-wave radar to provide a reference for real-time adjustment of dialysis parameters.

[0006] In a first aspect, embodiments of this application provide a method for adjusting hemodialysis parameters, the method comprising:

[0007] Millimeter waves are emitted toward the blood vessels and heart of the target object, and the echo signals of the millimeter waves are received;

[0008] Based on the echo signal, a first signal corresponding to a first point reflecting the vibration of the blood vessel and a second signal corresponding to a target area reflecting the vibration of the heart are obtained from the echo signal.

[0009] The first signal, the second signal, and the physiological information of the target object are input into a signal processing model so that the signal processing model can determine the physiological state information of the target object at the current moment, the physiological state information including hemodynamic characteristics and cardiac function characteristics;

[0010] The physiological state information, the current dialysis parameters of the target object, the historical dialysis parameters of the target object, and the corresponding historical physiological state information are input into the dialysis parameter prediction model so that the dialysis parameter prediction model outputs the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

[0011] Secondly, embodiments of this application provide a millimeter-wave radar, the millimeter-wave radar comprising:

[0012] Antenna array for transmitting millimeter waves toward the blood vessels and heart of a target object and receiving the echo signals of the millimeter waves;

[0013] The processor is configured to, based on the echo signal, acquire a first signal corresponding to a first point reflecting the vibration of the blood vessel, and a second signal corresponding to a target region reflecting the vibration of the heart;

[0014] The first signal, the second signal, and the physiological information of the target object are input into a signal processing model so that the signal processing model can determine the physiological state information of the target object at the current moment, the physiological state information including hemodynamic characteristics and cardiac function characteristics;

[0015] The physiological state information, the current dialysis parameters of the target object, the historical dialysis parameters of the target object, and the corresponding historical physiological state information are input into the dialysis parameter prediction model so that the dialysis parameter prediction model outputs the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

[0016] Thirdly, embodiments of this application provide a non-transitory machine-readable storage medium storing executable code, which, when executed by a processor of an electronic device, enables the processor to at least implement the method for adjusting hemodialysis parameters as described in the first aspect.

[0017] Fourthly, embodiments of this application provide a computer program product, which includes a computer program that, when executed by a processor, can implement the method for adjusting hemodialysis parameters as described in the first aspect.

[0018] In the hemodialysis parameter adjustment scheme provided in this application embodiment, if it is necessary to adjust the dialysis parameters of the target object during the dialysis process, the vibration of the target object's blood vessels and heart is measured by millimeter wave. The first signal generated by skin vibration caused by blood vessels and the second signal generated by skin vibration caused by the heart are obtained from the echo signal. The target object's physiological information, the first signal, and the second signal are processed by a signal processing model to obtain the target object's current physiological state information. Based on the current physiological state information, the target object's historical dialysis parameters, and corresponding historical physiological state information, the predicted dialysis parameters and / or predicted changes in dialysis parameters for the next moment are obtained through a dialysis parameter prediction model. This allows for the adjustment of the dialysis parameters of the target object's dialysis equipment. In a non-contact manner, the prediction of the target object's hemodialysis parameters is achieved, enabling timely adjustment of the dialysis parameters, making it more convenient and comfortable, and reducing the target object's discomfort. Furthermore, it can capture minute vibrations of blood vessels and heartbeats with high precision, and considers the influence of individual differences on the prediction results of hemodialysis parameters, ensuring that the adjusted dialysis parameters conform to the individual's situation, improving the accuracy of hemodialysis parameter prediction, and thus improving the accuracy of adjustment. The device is compact and easy to operate, enabling long-term continuous prediction and adjustment, and is suitable for various scenarios such as home dialysis and remote dialysis. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic diagram illustrating a scenario for adjusting hemodialysis parameters, provided as an embodiment of this application;

[0021] Figure 2 A flowchart illustrating a method for adjusting hemodialysis parameters provided in this application embodiment;

[0022] Figure 3 This is a schematic diagram of the structure of an electronic device provided in this embodiment. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. In addition, the timing of the steps in the following method embodiments is only an example and not a strict limitation.

[0024] It should be noted that, in the cases involving user information in the embodiments of this application, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse. In addition, the various models involved in this application (including but not limited to large language models or other models) comply with relevant laws and standards.

[0025] First, the terms or concepts involved in the embodiments of this application will be explained:

[0026] Millimeter waves are electromagnetic waves with wavelengths of 1-10 millimeters. The frequency range of millimeter waves is 30-300 GHz, falling within the wavelength range where microwaves and far-infrared waves overlap, thus exhibiting characteristics of both wavelengths.

[0027] Millimeter-wave radar is a technology that uses electromagnetic waves in the millimeter-wave frequency band (30-300 GHz) for target detection, ranging, and imaging. It analyzes information such as the target's distance, speed, and trajectory by emitting millimeter-wave signals and receiving reflected signals from the target.

[0028] Hemodialysis is a renal replacement therapy for end-stage renal disease and acute or chronic renal failure. It removes metabolic waste products (urea, creatinine, etc.) and excess water from the patient's blood through diffusion, convection, and ultrafiltration via a semipermeable membrane medium, while maintaining the body's electrolyte and acid-base balance. In this application, hemodialysis is simply referred to as dialysis. Dialysis is a vital means of sustaining life for patients with impaired kidney function. However, its effectiveness depends on strict monitoring during dialysis and timely management of complications.

[0029] During hemodialysis, parameters that need to be dynamically adjusted may include, but are not limited to, the following:

[0030] Blood flow rate (Qb): refers to the volume of blood flowing through an organ, tissue, or blood vessel per unit time, usually measured in milliliters per minute (mL / min). Higher blood flow rates result in higher solute clearance efficiency (such as urea clearance), but excessively high blood flow rates can lead to recirculation in the vascular system or increase the risk of sudden hypotension.

[0031] Dialysis fluid flow rate (Qd): Matched to blood flow rate, it affects the diffusion clearance of small molecule toxins (such as urea and creatinine).

[0032] Target Undigested Volume (UFV): This is set based on the difference between the patient's dry weight and actual weight, and needs to be dynamically adjusted based on bioimpedance analysis data.

[0033] Sodium ion concentration in dialysate: The normal range is usually 135-145 mmol / L. A high concentration can maintain blood volume stability but can cause thirst; a low concentration can increase the risk of hypotension.

[0034] Dialysis fluid temperature: Generally needs to be maintained within the range of 35.5-37.5℃.

[0035] Electrolyte balance: Adjust potassium (dialysis fluid potassium 2-4 mmol / L) and calcium (1.25-1.75 mmol / L) concentrations to avoid arrhythmias or muscle cramps during dialysis.

[0036] The following issues need to be considered during dialysis:

[0037] Intradialytic hypotension: the most common acute complication during dialysis, mainly caused by insufficient blood volume or autonomic nervous system regulation disorders due to rapid ultrafiltration. It often manifests as dizziness, weakness, and nausea, and in severe cases, loss of consciousness or arrhythmia may occur.

[0038] Dialysis Disequilibrium Syndrome: Caused by excessively rapid solute clearance, which disrupts the osmotic pressure of brain cells, leading to headaches, vomiting, altered consciousness, and even seizures.

[0039] Cardiac arrhythmias: Fluctuations in electrolytes (changes in potassium and calcium ion levels) and drastic fluctuations in blood volume can induce cardiac arrhythmias.

[0040] Arteriovenous fistula occlusion: Occlusion or stenosis of an arteriovenous fistula can lead to reduced blood flow, inadequate dialysis, and directly affect dialysis efficiency.

[0041] Cardiovascular Complications: Fluid changes and electrolyte imbalances during dialysis can increase the burden on the heart, leading to cardiovascular and cerebrovascular diseases such as heart failure, coronary heart disease, and hypertension.

[0042] Other indicators involved in the dialysis process:

[0043] Hemodynamics refers to the mechanics of blood flow in the cardiovascular system, primarily studying blood flow, blood flow resistance, blood pressure, and the relationships between them. This application refers to the parameters involved in hemodynamics as hemodynamic parameters.

[0044] Blood flow velocity (BFV) is a core parameter in hemodynamics, reflecting the speed and characteristics of blood flow in blood vessels. Its changes directly reflect the functional state of the cardiovascular system and are closely related to physiological indicators such as blood pressure, heart rate, and vascular elasticity. With the development of clinical medicine, accurate and continuous acquisition of blood flow velocity has become a key technological path for realizing intelligent health monitoring and personalized medicine.

[0045] The main indicators of blood flow velocity include:

[0046] Peak systolic velocity (PS): This reflects the maximum velocity of blood flow during cardiac systole and can be used to classify the degree of arterial stenosis and assess the speed of blood flow through the valves.

[0047] End-diastolic velocity (ED): This reflects the residual velocity of blood flow during cardiac diastole. It is related to vascular resistance and can be used to assess arterial stenosis and peripheral vascular disease.

[0048] Time-averaged maximum velocity (TAmax): This is the time-averaged maximum blood flow velocity over the entire cardiac cycle. It can be used to assess the trend of maximum blood flow and is suitable for examining arterial stenosis, valvular disease, and hemodynamic abnormalities.

[0049] Time-mean velocity (TAmean): This is the average blood flow velocity over the entire cardiac cycle. It reflects the average hemodynamic state and is related to cardiac output.

[0050] Time-velocity integral (VTI): The integral of blood flow velocity over time, which reflects the total velocity path of blood flow through a certain cross-section of a blood vessel during a cardiac cycle.

[0051] In addition to blood flow velocity, hemodynamic parameters also include, but are not limited to:

[0052] Volume flow rate refers to the volume of blood passing through a cross-section of a blood vessel per unit time, usually expressed in mL / s. Decreased flow rate suggests insufficient cerebral perfusion, while high flow rate may be accompanied by arteriovenous fistula or abnormal collateral blood supply; postoperative flow rate changes are used to assess treatment efficacy.

[0053] Pulsatility Index (PI): This measure is used to assess the amplitude of blood flow waveforms. A high PI may indicate increased distal resistance, such as small vessel disease or intracranial hypertension; a low PI may be seen in arterial dilation or aortic regurgitation.

[0054] Vascular compliance is the volumetric response of blood vessels to changes in pressure, reflecting vascular elasticity. Decreased compliance in arteriosclerosis leads to increased pulse pressure; it is a key physiological indicator of arterial aging, stiffness, and cardiovascular risk.

[0055] Wall shear stress (WSS): The tangential stress exerted by blood flow on the inner wall of a blood vessel, measured in Pa. Areas with long-term low shear stress are prone to atherosclerotic plaque formation; high WSS may induce plaque rupture and participate in arterial remodeling and inflammatory responses.

[0056] Pulse wave velocity (PWV) is the speed at which a pulse wave travels through an artery and is considered the "gold standard" parameter for reflecting arterial stiffness. Elevated PWV is closely related to arteriosclerosis and elevated blood pressure; it is also independently positively correlated with the risk of cardiovascular and cerebrovascular events such as stroke and heart failure, making it an important indicator for predicting prognosis.

[0057] Cardiac function indicators may include, but are not limited to, the following:

[0058] Arterial blood pressure (BP): Arterial blood pressure is the pressure exerted by blood against the walls of arteries as it flows through them. It is usually divided into systolic pressure and diastolic pressure. Changes in blood pressure directly affect the efficiency of blood flow. Both high and low blood pressure can affect the workload on the heart and the health of blood circulation.

[0059] Vascular resistance is the resistance that blood encounters when flowing through blood vessels. An increase in vascular resistance may indicate peripheral vasoconstriction or arteriosclerosis, and is a marker of hypertension and atherosclerosis.

[0060] Pulse wave velocity (PWV) is the speed at which a pulse wave travels through an artery and is an important indicator of vascular elasticity. An increase in PWV suggests arteriosclerosis, reflects the rigidity of blood vessels, and is generally closely related to the risk of cardiovascular disease.

[0061] Cardiac output (CO): Cardiac output is the total amount of blood pumped by the heart per minute. It is a key indicator of the heart's pumping capacity and reflects its overall function. A decrease in cardiac output may indicate heart failure or other heart diseases.

[0062] Ejection fraction (EF): Ejection fraction refers to the percentage of blood ejected from the left ventricle during each cardiac contraction, relative to the total blood volume of the left ventricle. EF is an important indicator for assessing cardiac pumping function. Low EF values ​​are commonly seen in left ventricular failure or cardiac damage following myocardial infarction.

[0063] During hemodialysis, patients face complex changes in body fluids, electrolytes, and hemodynamics. To ensure dialysis safety, it is necessary to dynamically adjust dialysis parameters.

[0064] The current method of adjusting dialysis parameters usually involves professionals intermittently obtaining the patient's blood pressure, heart rate, and other data. However, this method is slow to respond to sudden hypotension or arrhythmia and is only suitable for periodic assessments, making it difficult to use for real-time control.

[0065] The above methods require professional operation and cannot be applied to various scenarios such as home dialysis and remote dialysis. They also cannot monitor the dialysis process in real time. Therefore, they cannot adjust dialysis parameters in real time according to the current dialysis situation, and in particular, they cannot detect high-risk events such as sudden hypotension, arrhythmia or blood flow blockage in a timely manner, resulting in a significant response lag.

[0066] To overcome these shortcomings, this application proposes a hemodialysis parameter adjustment scheme applicable to various scenarios, capable of providing real-time and continuous adjustment information. The hemodialysis parameter adjustment scheme provided in this application is described in detail below.

[0067] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating a scenario for adjusting hemodialysis parameters, provided as an embodiment of this application. Figure 1As shown, the hemodialysis parameter adjustment scenario includes a hemodialysis parameter adjustment device 101. The hemodialysis parameter adjustment device 101 is based on millimeter-wave radar; hereinafter, it may also be referred to as millimeter-wave radar. The hemodialysis parameter adjustment device 101 may include a millimeter-wave transmitting / receiving module and a processor. The millimeter-wave transmitting / receiving module and the processor are communicatively connected. Typically, the millimeter-wave transmitting / receiving module includes an antenna array for transmitting millimeter-wave signals and receiving the echo signals. The processor processes the echo signals to obtain the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

[0068] In practical applications, if it is necessary to adjust dialysis parameters during the dialysis process of a target subject, the hemodialysis parameter adjustment device 101 can be directed towards the target subject's blood vessels and heart to emit millimeter waves. The target subject can be a human or an animal. Since millimeter waves have a certain penetrating power and can penetrate non-metallic objects such as clothing, the hemodialysis parameter adjustment device 101 can be set within a preset distance range from the target subject. For example, the preset distance range can be within 90 centimeters of the target subject. For example, the hemodialysis parameter adjustment device 101 can also be placed on the target subject's skin as a patch. To obtain more accurate evaluation results, no metal objects should be placed between the hemodialysis parameter adjustment device 101 and the target subject. Generally, apart from the clothing worn by the target subject, no other objects should be placed between the hemodialysis parameter adjustment device 101 and the target subject. The hemodialysis parameter adjustment device 101 typically emits millimeter waves simultaneously towards the target subject's blood vessels and heart, thus allowing the same device to simultaneously obtain the target subject's hemodynamic characteristics and cardiac function characteristics.

[0069] Because the pulsation of blood vessels and the beating of the heart in the target object will cause vibrations in the surrounding tissues, these vibrations are relatively small and can be called micro-movements. Therefore, the micro-movements of superficial blood vessels and the beating of the heart can be determined by observing these micro-movements in the skin around the heart. After setting the adjustment device 101 for hemodialysis parameters, it needs to be positioned so that it emits millimeter waves towards the target object. The millimeter waves are reflected after passing through the skin of the target object, and the adjustment device 101 can receive the reflected echo signals. Taking a human as an example, the human carotid artery is close to the skin, and its pulsation is clearly visible on the skin and easy to measure. Moreover, hemodynamic parameters can usually be obtained from the blood flow in the carotid artery. In addition, the beating of the heart will also cause vibrations in the surrounding skin. Therefore, the adjustment device 101 for hemodialysis parameters can emit millimeter waves towards the location of the target object's carotid artery and heart.

[0070] The hemodialysis parameter adjustment device 101 processes the echo signal as follows: Based on the echo signal, it obtains a first signal corresponding to a first point reflecting the vibration of the blood vessels and a second signal corresponding to a target area reflecting the vibration of the heart. The first signal, the second signal, and the physiological information of the target object are input into a signal processing model to determine the physiological state information of the target object at the current moment. This physiological state information includes hemodynamic characteristics and cardiac function characteristics. The physiological state information, the current dialysis parameters of the target object, the historical dialysis parameters of the target object, and the corresponding historical physiological state information are input into a dialysis parameter prediction model to output the predicted dialysis parameters and / or the predicted changes in dialysis parameters for the next moment.

[0071] Furthermore, the hemodialysis parameter adjustment device 101 can also communicate with the computer device 102 via wired or wireless means. Considering the predicted dialysis parameters and / or predicted changes in dialysis parameters output by the hemodialysis parameter adjustment device 101 for the next moment, this information can be transmitted to the computer device 102 used by professionals, allowing them to know the predicted dialysis parameters and / or predicted changes in dialysis parameters for the next moment, and to adjust the dialysis parameters accordingly. Optionally, the computer device 102 can be connected to a display to show the predicted dialysis parameters and / or predicted changes in dialysis parameters obtained by the hemodialysis parameter adjustment device 101.

[0072] Furthermore, the hemodialysis parameter adjustment device 101 can also communicate with the dialysis device 103 via wired or wireless means, and adjust the dialysis parameters of the dialysis device based on the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment. For example, the dialysis device 103 can be a hemodialysis machine.

[0073] Based on the above scenario, the hemodialysis parameter adjustment method provided in this application, if it is necessary to adjust the dialysis parameters of the target object during the dialysis process, measures the vibration of the target object's blood vessels and heart using millimeter waves to obtain a first signal generated by skin vibration caused by blood vessels in the echo signal, and a second signal generated by skin vibration caused by the heart in the echo signal. The target object's physiological information, the first signal, and the second signal are processed by a signal processing model to obtain the target object's current physiological state information. Based on the current physiological state information, as well as the target object's historical dialysis parameters and corresponding historical physiological state information, a dialysis parameter prediction model is used to obtain the predicted dialysis parameters and / or the predicted changes in dialysis parameters for the next moment, thereby allowing adjustment of the dialysis parameters of the target object's dialysis equipment. This method enables the prediction of the target object's hemodialysis parameters in a non-contact manner, allowing for timely adjustment of dialysis parameters, making the process more convenient and comfortable, and reducing the target object's discomfort. Furthermore, it can accurately capture minute vibrations in blood vessels and heartbeats, and takes into account the impact of individual differences on the prediction results of hemodialysis parameters. This allows the adjusted dialysis parameters to be tailored to the individual's situation, improving the accuracy of hemodialysis parameter prediction and thus improving the accuracy of adjustment. The device is compact, easy to operate, and can achieve long-term continuous prediction and adjustment, making it suitable for various scenarios such as home dialysis and remote dialysis.

[0074] The following describes in detail the execution process of the hemodialysis parameter adjustment method provided in the embodiments of this application, with reference to the accompanying drawings. Optionally, this method can be derived from the above... Figure 1 The device 101, which adjusts the hemodialysis parameters shown, is activated.

[0075] Figure 2 A flowchart of a method for adjusting hemodialysis parameters provided in this application embodiment is shown below. Figure 2 As shown, the method includes the following steps:

[0076] 201. Transmit millimeter waves to the blood vessels and heart of the target object and receive the echo signals of the millimeter waves.

[0077] 202. Based on the echo signal, obtain the first signal corresponding to the first point reflecting the vibration of the blood vessel, and the second signal corresponding to the target area reflecting the vibration of the heart.

[0078] 203. Input the first signal, the second signal, and the physiological information of the target object into the signal processing model so that the signal processing model can determine the physiological state information of the target object at the current moment. The physiological state information includes hemodynamic characteristics and cardiac function characteristics.

[0079] 204. Input the physiological state information, the current dialysis parameters of the target object, the historical dialysis parameters of the target object, and the corresponding historical physiological state information into the dialysis parameter prediction model so that the dialysis parameter prediction model outputs the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

[0080] In this embodiment, if it is necessary to predict the dialysis parameters of the target object, the hemodialysis parameter adjustment device simultaneously transmits millimeter waves to the target object's blood vessels and heart, and receives the echo signals of the millimeter waves.

[0081] Based on the echo signal, a first signal corresponding to a first point reflecting the vibration of blood vessels and a second signal corresponding to a target area reflecting the vibration of the heart are obtained from the echo signal. Since millimeter waves are reflected after touching the skin of the target object, the location of the target object's skin can be determined based on the echo signal. Since the slight vibration of the skin is driven by superficial blood vessels and the heart, the first point of skin vibration driven by superficial blood vessels and the target area of ​​skin vibration driven by heartbeat can be determined. The first point can be considered as a point reflecting the vibration of blood vessels in the target object, for example, a point on the skin at the carotid artery of the target object. The target area can be considered as a point reflecting the heartbeat of the target object, for example, a point on the skin at the heart of the target object. The signal component reflected from the first point is obtained as the first signal, and the signal component reflected from the target area is obtained as the second signal.

[0082] Optionally, taking the carotid artery as an example, if the hemodialysis parameter adjustment device is placed directly facing the target object, then in the coordinate system with the hemodialysis parameter adjustment device as the origin, the carotid artery is located above the heart. Since the location of the target object can be located through the echo signal, the first point with the strongest intensity in the upper half of the echo signal can be determined to be the skin location around the carotid artery, and the second point with the strongest intensity in the lower half of the echo signal can be determined to be the skin location around the left ventricle of the heart. The location of the heart can then be determined from the location of the left ventricle, that is, the target area corresponding to the heart can be determined from the location of the second point.

[0083] Then, the first signal, the second signal, and the physiological information of the target object are input into the signal processing model so that the signal processing model can determine and output the physiological state information of the target object at the current moment based on the first signal, the second signal, and the physiological information of the target object.

[0084] Due to individual differences, the physiological state information obtained for the same hemodynamic and cardiac function characteristics may differ among different individuals. That is, individual differences have a certain impact on the obtained hemodynamic and cardiac function characteristics. Therefore, this application needs to obtain the physiological information of the target object and evaluate the hemodynamic indicators of the target object through the physiological information of the target object and the echo signal generated by the target object's blood vessels.

[0085] Optionally, the above physiological information may include, but is not limited to, gender, age, height, weight, and blood pressure.

[0086] Furthermore, physiological information may also include dialysis age.

[0087] The signal processing model described above is a pre-built and trained model. It is used to determine and output corresponding physiological state information based on the input signal and physiological information.

[0088] Optionally, the signal processing model is obtained by training the initial signal processing model using a first training sample set, and the loss value obtained during training satisfies a preset convergence condition. The first training sample set contains multiple first samples, and each of the multiple first samples contains a corresponding first signal, a second signal, and physiological state information of the sample.

[0089] Furthermore, the signal processing model can be based on an encoder-decoder network. An encoder-decoder network consists of an encoder and a decoder connected sequentially. The encoder maps the input signal and physiological information (which can be referred to as temporal data) to a latent space with high-level abstract features, outputting high-dimensional physiological state information. This high-dimensional physiological state information refers to the high-dimensional features composed of mechanical parameters such as the amplitude, velocity, and acceleration of vascular pulsation. This high-dimensional physiological state information is then fed into the decoder, which extracts hemodynamic and cardiac function features. Specifically, the sequence data input to the encoder is processed through a convolutional neural network to extract local temporal features. After convolution, an activation function (ReLU) and a max-pooling layer are applied to enhance the network's nonlinear expressive power and model training stability. The decoder is used to reduce the dimensionality of the extracted features, ultimately outputting the physiological state information.

[0090] The physiological state information output by the aforementioned signal processing model is used to characterize the blood vessels and heart of the target object at the current moment. This physiological state information can include hemodynamic characteristics and cardiac function characteristics.

[0091] Optional hemodynamic characteristics include blood flow velocity parameters. Blood flow velocity parameters include maximum blood flow velocity, minimum blood flow velocity, and blood flow velocity time-velocity integral.

[0092] Furthermore, blood flow velocity parameters may also include at least one of the following: peak systolic velocity, end-diastolic velocity, time-averaged maximum velocity, and time-averaged velocity.

[0093] Furthermore, hemodynamic characteristics may also include at least one of volumetric flow rate, pulsatility index, vascular compliance, wall shear stress, and pulse wave velocity.

[0094] Optionally, cardiac function characteristics may include: ejection fraction, pulse wave velocity, systolic blood pressure, and diastolic blood pressure. Among these, the ejection fraction is a parameter that can be used to equivalently derive cardiac contractility.

[0095] Next, since we have obtained the physiological state information of the target object at the current moment, which is equivalent to understanding the target object's physical adaptation during the current dialysis, we can use the target object's physiological state information at the current moment, combined with the dialysis parameters at the current moment, as well as the dialysis parameters and corresponding physiological state information of the target object during previous dialysis sessions, to predict the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment through the dialysis parameter prediction model.

[0096] Dialysis parameters refer to the control parameters during the dialysis process. Dialysis parameters may include, but are not limited to: dialysate temperature, dialysate flow rate, target water removal volume, and dialysate sodium ion concentration.

[0097] The aforementioned dialysis parameter prediction model is a pre-established and trained model. It is used to determine and output predicted dialysis parameters and / or predicted changes in dialysis parameters based on physiological state information, the target subject's current dialysis parameters, the target subject's historical dialysis parameters, and corresponding historical physiological state information.

[0098] Optionally, the dialysis parameter prediction model is trained on the initial dialysis parameter prediction model using a second training sample set, and is obtained when the loss value obtained during training satisfies a preset convergence condition. The second training sample set contains multiple second samples, each of which contains corresponding physiological state information of the sample, the dialysis parameters at the current moment, the historical dialysis parameters of the sample, historical physiological state information of the sample, and a label. The label can be the predicted dialysis parameters of the sample at the next moment and / or the change in the predicted dialysis parameters of the sample.

[0099] The above-mentioned dialysis parameter prediction model is based on a Markov model and was obtained after training and convergence.

[0100] Optionally, the dialysis parameter prediction model is used to determine the probability distribution of the current first circulatory cardiac function state and the predicted circulatory cardiac function state at the next moment, and to obtain the maximum a posteriori estimated circulatory cardiac function state. Based on the current first circulatory cardiac function state and the maximum a posteriori estimated circulatory cardiac function state, the predicted change in dialysis parameters at the next moment is obtained.

[0101] In summary, the solution provided in the above embodiments of this application transmits and receives millimeter-wave echo signals to and from the blood vessels and heart of the target object. Based on the echo signals, a first signal corresponding to a first point reflecting the vibration of the blood vessels and a second signal corresponding to the target area reflecting the vibration of the heart are obtained. The first signal, the second signal, and the physiological information of the target object are input into a signal processing model to obtain the physiological state information of the target object at the current moment, including hemodynamic characteristics and cardiac function characteristics. Based on the physiological state information at the current moment, and the target object's historical dialysis parameters and corresponding historical physiological state information, the predicted dialysis parameters and / or predicted changes in dialysis parameters at the next moment are obtained through a dialysis parameter prediction model, thereby allowing adjustment of the dialysis parameters of the dialysis device for the target object. Because millimeter waves can penetrate clothing or other non-metallic materials and have strong penetrability, the physiological state of the target object can be monitored without contact, enabling the prediction of the target object's hemodialysis parameters and timely adjustment of dialysis parameters, making it more convenient and comfortable and reducing discomfort for the subject. Furthermore, millimeter waves offer high distance and velocity resolution, enabling precise capture of minute vibrations in blood vessels and heartbeats, thus improving the accuracy of hemodialysis parameter prediction and consequently, the accuracy of regulation. By incorporating the physiological information of the target individual and considering the impact of individual differences on physiological state information, the obtained physiological state information is more accurate, further enhancing the accuracy of hemodialysis parameter prediction and regulation. The device is compact, easy to operate, and capable of long-term continuous prediction and regulation, making it suitable for various scenarios such as home dialysis and remote dialysis.

[0102] In some embodiments, in step 202, a first signal corresponding to a first point reflecting the vibration of the blood vessel and a second signal corresponding to a target area reflecting the vibration of the heart can be determined based on radar principles for determining the position and velocity of the target object. For example, the blood vessel includes the carotid artery, and one possible implementation of step 202 is:

[0103] 2021. Perform signal processing on the echo signal to obtain the intermediate frequency signal corresponding to the echo signal.

[0104] 2022. Determine the spatial orientation spectrum of the target object based on the intermediate frequency signal.

[0105] 2023. Based on the relative positional relationship between the carotid artery and the heart, the spatial orientation spectrum is divided into a first spatial orientation spectrum corresponding to a first region and a second spatial orientation spectrum corresponding to a second region; the first region contains the carotid artery and the second region contains the heart.

[0106] 2024. Based on the first spatial orientation spectrum, determine the first point with the greatest intensity in the first spatial orientation spectrum, and determine the location information of the first point.

[0107] 2025. Based on the location information of the first point, extract the first signal corresponding to the first point from the intermediate frequency signal.

[0108] 2026. Based on the second spatial orientation spectrum, determine the second point with the strongest intensity in the second spatial orientation spectrum, and determine the location information of the second point.

[0109] 2027. Based on the location information of the second point, determine the location information of the target area reflecting the vibration of the heart.

[0110] 2028. Based on the location information of the target area, extract the second signal corresponding to the target area from the intermediate frequency signal.

[0111] In this embodiment, based on radar principles, signal processing is performed on the echo signal to obtain the intermediate frequency (IF) signal corresponding to the echo signal. The spatial orientation spectrum of the target object is then determined based on the IF signal.

[0112] In an optional embodiment, step 2022 can be implemented by the following steps 20221-20223:

[0113] Step 20221: Determine the distance range of each distance chamber based on the intermediate frequency signal.

[0114] Furthermore, the process of determining the distance range of each warehouse is as follows:

[0115] The frequency information of the intermediate frequency signal and the distance to the measured target have a linear relationship as shown in the following formula (1):

[0116]

[0117] Among them, f IF τ is the frequency of the intermediate frequency signal, S is the sweep frequency of the millimeter-wave radar, τ is the round-trip time of the millimeter-wave radar and the target, R is the distance between the millimeter-wave radar and the target, and c is the speed of millimeter-wave propagation in the air.

[0118] In actual measurements, there are many reflective objects in the environment, which leads to f IF There are multiple frequency domain components. To extract the reflection signals related to superficial blood vessels and eliminate interfering objects at other distances, it is necessary to...IF The system is divided into different distance bins using a fast Fourier transform (FFT).

[0119]

[0120] Among them, f bin This refers to the intermediate frequency (IF) corresponding to the range chamber; samples are the radar's sampling points; samplesIndex is the sampling point identifier; f s It is the radar's sampling frequency, which is f in the above formula (1). IF .

[0121] Based on the above formulas (1) and (2), the distance range represented by each distance bin can be determined. bin As shown in formula (3) below:

[0122]

[0123] Among them, f bin is the intermediate frequency corresponding to the range chamber, c is the speed of propagation of millimeter waves in the air, and S is the sweep frequency of the millimeter-wave radar.

[0124] Step 20222: Determine the distance bin to which the target object belongs based on the preset distance range between the target object and the target object.

[0125] Objects in physical space will generate millimeter-wave echo signals, so it is necessary to eliminate interference from the echo signals of objects other than the target object.

[0126] The preset distance range refers to the distance range between the hemodynamic assessment device and the target object. For example, the distance between the hemodynamic assessment device and the target object is generally set within 90 centimeters, so the preset distance range is within 90 centimeters. Distance chambers within this preset distance range can be identified as the distance chambers to which the target object belongs.

[0127] Step 20223: Based on beamforming, determine the spatial orientation spectrum within the distance compartment to which the target object belongs.

[0128] For hemodynamic assessment devices, the relative positions of multiple antennas in the antenna array are fixed. Beamforming processing is performed on the signals received from these antennas, and the echo signals are mapped into three-dimensional space to obtain the target object's location information. This target object's location information can be three-dimensional coordinates.

[0129] Optionally, assuming the antenna array contains M antenna elements, the echo signal collected by each antenna element is as shown in formula (4).

[0130]

[0131] Among them, X m (k) is the echo signal received by the m-th antenna, τ m R(k) is the propagation delay difference between the antenna element and the target object, k is the range cell index, R(k) is the frequency domain signal of different range cells, k is the range-dimensional spectrum signal, and f is the range-dimensional spectrum signal. c It is the carrier frequency of millimeter-wave radar.

[0132] By performing beamforming processing on the signals collected by M antenna elements, the location information of the target object can be obtained.

[0133] Based on the location information of the target object, the signal components of each point on the target object can be extracted.

[0134] Optionally, the signal component of each point on the target object can be determined by the following formula (5).

[0135]

[0136] Where n is the number of antenna elements in the antenna array, x i and y i Here, represents the spatial coordinates of the i-th antenna, Bin(z) is the signal component at distance z, and λ is the wavelength of the millimeter wave. This represents the phase information extracted from the echo signal.

[0137] The signal component at each point is converted into a phase signal through Fourier transform, yielding the power spectral density at each point. This power spectral density can also be referred to as the spatial orientation spectrum.

[0138] Optionally, the spatial orientation spectrum can be represented by the following formula (6).

[0139]

[0140] Where P(θ,φ,k) is the spatial spectral intensity when the spatial position is azimuth angle φ, elevation angle θ, and range k, X m (k) is the echo signal received by the m-th antenna, M is the number of antenna elements in the antenna array, and d m λ is the spatial distance between the m-th antenna element and the reference element, and λ is the wavelength of the millimeter wave.

[0141] After obtaining the spatial orientation spectrum, since the carotid artery is located above the heart relative to the millimeter-wave radar, the obtained spatial orientation spectrum can be divided into two spatial orientation spectra: the upper spatial orientation spectrum corresponds to the first region containing the carotid artery, and the lower spatial orientation spectrum corresponds to the second region containing the heart.

[0142] The first and second spatial azimuth spectra are processed separately. When processing the first spatial azimuth spectrum, the first point with the highest intensity in the first spatial azimuth spectrum can be determined, and its location information can be identified. Based on the location information of the first point, the first signal corresponding to the first point is extracted from the intermediate frequency signal. For example, the location information can be spatial coordinates.

[0143] When processing the second spatial orientation spectrum, the second point with the strongest intensity in the spectrum can be identified, and its location information can be determined. This second point reflects the beating status of the left ventricle. After determining the location of the left ventricle, based on the heart structure, the locations of the other atria and ventricles (excluding the left ventricle) can be estimated, i.e., the location of the heart can be determined. In other words, based on the location information of the second point, the location information of the target area of ​​the skin reflecting the heart's vibration can be determined.

[0144] Optionally, when processing the second spatial orientation spectrum, the second point can be obtained as follows: Based on the second spatial orientation spectrum, determine the signal energy spectrum within the heart rate frequency band. Obtain the second point with the highest intensity in the signal energy spectrum.

[0145] Furthermore, the power spectral density can be denoted as PSD(x,y,z,f), which represents the signal component at frequency f in the (x,y,z) coordinate system. The unit of f is (beats / minute), and this frequency is the phase change frequency, which is the frequency at which the chest cavity undergoes minute displacements. Since the heart rate frequency range is 40-200 beats / minute, the sum of the power spectral densities at frequencies between 40 and 200 can be obtained. This sum can be called the heartbeat signal, denoted by F. heart (x,y,z) represents F heart (x,y,z) can be represented by the following formula (7):

[0146]

[0147] F heart The higher the value of (x, y, z), the higher the energy of the heartbeat in the signal at that coordinate, and the closer it is to the left ventricle. Therefore, F can be obtained. heart Find the second point corresponding to the maximum value in (x,y,z) and obtain the coordinate information of the second point.

[0148] Furthermore, the coordinate information (x0, y0, z0) of the second point is obtained using the following formula (8):

[0149] (x0, y0, z0) = argmax (x,y,z) F heart (x,y,z) Formula (8)

[0150] Where, argmax (x,y,z) This represents the value of the variable (x, y, z) that makes the function reach its maximum value.

[0151] Based on the location information of the target area, the second signal corresponding to the target area is extracted from the intermediate frequency signal.

[0152] In this embodiment, based on radar principles, a first signal reflecting skin vibrations of superficial blood vessels in the target object and a second signal reflecting skin vibrations of the target object's heart can be acquired. The first and second signals, along with the target object's physiological information, are input into a signal processing model to obtain the target object's physiological state information at the current moment, including hemodynamic and cardiac function characteristics. Based on the current physiological state information, the target object's historical dialysis parameters, and corresponding historical physiological state information, a dialysis parameter prediction model is used to obtain the predicted dialysis parameters and / or predicted changes in dialysis parameters for the next moment, thereby allowing adjustment of the dialysis parameters of the target object's dialysis equipment. Because millimeter waves can penetrate clothing or other non-metallic materials, they have strong penetrability, enabling non-contact monitoring of the target object's physiological state and prediction of its hemodialysis parameters. This allows for timely adjustment of dialysis parameters, making the process more convenient and comfortable, and reducing discomfort for the subject. Furthermore, millimeter waves have high distance and velocity resolution, enabling high-precision capture of minute vascular vibrations and heartbeats, improving the accuracy of hemodialysis parameter prediction and thus improving adjustment accuracy. By incorporating the physiological information of the target individual and taking into account the impact of individual differences on physiological state information, the obtained physiological state information is more accurate, thereby improving the accuracy of hemodialysis parameter prediction and thus improving the accuracy of regulation. The device is compact, easy to operate, and can achieve long-term continuous prediction and regulation, making it suitable for various scenarios such as home dialysis and remote dialysis.

[0153] In some embodiments, the echo signal contains noise signals caused by factors other than superficial vascular pulsation and skin vibration caused by heartbeat. Before inputting the first and second signals into the signal processing model, the first and second signals can be preprocessed separately. This preprocessing is used to filter out noise signals caused by factors other than vascular pulsation and heartbeat in the signal components. Specifically, the first signal undergoes a first preprocessing to obtain a preprocessed first signal, which is used to filter out noise signals caused by factors other than vascular vibration in the target signal. The second signal undergoes a second preprocessing to obtain a preprocessed second signal; this second preprocessing is used to filter out noise signals caused by factors other than heartbeat in the target signal. Accordingly, in step 203, the signals input to the signal processing model are the preprocessed first signal, the preprocessed second signal, and the physiological information of the target object.

[0154] By preprocessing the first signal, noise signals caused by factors other than vascular vibration can be filtered out, thereby making the assessment results of hemodynamic parameters obtained from the preprocessed first signal more accurate.

[0155] In an optional embodiment, the first preprocessing includes at least one of a first filtering process, a first outlier processing process, and a first normalization process; the first filtering process is used to filter out noise in the first signal caused by respiratory displacement and / or human body movement; the outlier processing process is used to remove outliers from the first signal; and the normalization process is used to normalize the first signal according to a preset amplitude ratio.

[0156] Optionally, the specific implementation process of the first preprocessing is described below. Besides vascular micromotion information, the echo signal also contains noise such as respiratory displacement and human body movement. To extract a more accurate vascular pulsation phase, this noise needs to be filtered out. This interference mainly comes from respiratory effects; the target object's own breathing causes vibrations on the skin surface. Additionally, some dialysis patients experience interdialysis hypotension syndrome, which can also cause unnatural tremors. The vibrations of the skin surface and body caused by breathing are at a higher frequency than the vascular pulsation to be detected. Therefore, the first preprocessing can include a first filtering process. Specifically, the first signal is preprocessed to obtain a first preprocessed first signal, which can be as follows: the first signal is subjected to a first filtering process to obtain a filtered first signal component; the first filtering process is used to filter out noise caused by respiratory displacement and / or human body movement in the first signal. The signal component after the first filtering process is used as the first preprocessed first signal.

[0157] Furthermore, although the amplitude of micro-movement is smaller than that of respiration and body movement, the acceleration it causes is greater. The difference in acceleration between vascular micro-movement and respiration and body movement information can be used to perform differential acceleration filtering on the first signal to obtain the filtered first signal.

[0158] The first signal after filtering can be obtained according to the following formula (9).

[0159]

[0160] Among them, Diff_p i p is the first signal after filtering. i Let be the spatial spectrum at time i, and T be the time interval of the intermediate frequency signal.

[0161] In an optional embodiment, human motion or unstable background noise can interfere with millimeter-wave signals and affect the subsequent evaluation of hemodynamic parameters. To achieve hemodynamic parameter evaluation in real-world scenarios, noise interference must be properly handled to avoid misjudgment. Therefore, the first preprocessing may further include first outlier processing. After obtaining the filtered first signal, the preprocessing further includes: performing first outlier processing on the filtered first signal to obtain an outlier-processed first signal; the first outlier processing is used to remove outliers from the first signal. The outlier-processed first signal is then used as the preprocessed first signal.

[0162] For example, the first outlier handling can be implemented through the following process: The filtered first signal is divided into preset time windows. For example, it can be divided into continuous time windows of 200ms each (with a step size of 50ms). The maximum absolute amplitude (MAA) within each time window is calculated. When the MAA value of a certain time window exceeds a preset energy threshold, the first signal within that time window is considered an outlier, and the first signal within that time window is removed from the first signal set. Optionally, the preset energy threshold can be set to three times the average of multiple MAAs, or it can be set to other values; this application does not limit this.

[0163] In an optional embodiment, since the preprocessed first signal is to be input into the model for processing, it is necessary to normalize the signal before it is input into the model to conform to the model's input data format. Therefore, the first preprocessing may further include a first normalization process. After outlier handling, the outlier-handled first signal may be subjected to a first normalization process to obtain a normalized first signal. The normalization process is used to normalize the first signal according to a preset amplitude ratio. Accordingly, the normalized first signal is used as the first preprocessed first signal.

[0164] Furthermore, since the first normalization process cannot affect the vascular micromotion information contained in the first signal, the relative amplitudes and frequencies of the micromotion components in different cardiac cycles can be well preserved after the first normalization. Therefore, the first signal after outlier processing can be subjected to the first normalization process proportionally. For example, the first normalization process can be the first root mean square normalization process.

[0165] Accordingly, the second preprocessing includes at least one of a second filtering process, a second outlier processing, and a second normalization process. The second filtering process is used to filter out noise in the second signal caused by respiratory displacement and / or human body movement. The outlier processing is used to remove outliers from the second signal. The normalization process is used to normalize the second signal according to a preset amplitude ratio.

[0166] Optionally, the specific implementation process of the second preprocessing is described below. Besides heartbeat information, the echo signal also contains noise such as respiratory displacement and human body movement. To extract a more accurate heartbeat phase, this noise needs to be filtered out. This interference mainly comes from respiratory effects; the target object's own breathing causes vibrations on the skin surface. Additionally, some dialysis patients, such as those with interdialysis hypotension syndrome, may also experience unnatural body tremors. The vibrations of the skin surface and body caused by breathing are at a higher frequency than the heartbeat we want to detect. Therefore, the aforementioned second preprocessing can include a second filtering process. Specifically, the second signal is preprocessed to obtain a preprocessed second signal, which can be as follows: the second signal is subjected to a second filtering process to obtain a filtered second signal component; the second filtering process is used to filter out noise caused by respiratory displacement and / or human body movement in the second signal. The signal component after the second filtering process is used as the preprocessed second signal.

[0167] Furthermore, although the amplitude of micro-movements is smaller than that of breathing and body movements, the acceleration they cause is greater. The difference in acceleration between heartbeat and breathing / body movement information can be used to perform differential acceleration filtering on the second signal to obtain the filtered second signal.

[0168] The filtered second signal can be obtained according to the following formula (9).

[0169]

[0170] Among them, Diff_p i p is the second signal after filtering. i Let be the spatial spectrum at time i, and T be the time interval of the intermediate frequency signal.

[0171] In an optional embodiment, human motion or unstable background noise can interfere with millimeter-wave signals and affect the subsequent evaluation of hemodynamic parameters. To achieve hemodynamic parameter evaluation in real-world scenarios, noise interference must be properly handled to avoid misjudgment. Therefore, the second preprocessing may further include second outlier processing. After obtaining the filtered second signal, the preprocessing further includes: performing second outlier processing on the filtered second signal to obtain an outlier-processed second signal; the second outlier processing is used to remove outliers from the second signal. The outlier-processed second signal is then used as the preprocessed second signal.

[0172] For example, the second outlier handling can be implemented through the following process: The filtered second signal is divided into preset time windows. For example, it can be divided into continuous time windows of 200ms (50ms step). The maximum absolute amplitude (MAA) within each time window is calculated. When the MAA value of a certain time window exceeds a preset energy threshold, the second signal within that time window is considered an outlier, and the second signal within that time window is removed from the list of second signals. Optionally, the preset energy threshold can be set to three times the average of multiple MAAs, or it can be set to other values; this application does not limit this.

[0173] In an optional embodiment, since the preprocessed second signal is to be input into the model for processing, it is necessary to normalize the signal before it is input into the model to conform to the model's input data format. Therefore, the second preprocessing may further include a second normalization process. After outlier handling, the outlier-handled second signal may be subjected to a second normalization process to obtain a normalized second signal. The normalization process is used to normalize the second signal according to a preset amplitude ratio. Accordingly, the normalized second signal is used as the second preprocessed second signal.

[0174] Furthermore, since the second normalization process cannot affect the heartbeat information contained in the second signal, the relative amplitudes and frequencies of the micro-motion components within different cardiac cycles can be well preserved after the second normalization. Therefore, the second signal after outlier processing can be subjected to a second normalization process proportionally. For example, this second normalization process can be a second root mean square normalization process.

[0175] In some embodiments, step 205 may be included after step 204.

[0176] Step 205: Send the predicted dialysis parameters and / or predicted changes in dialysis parameters at the next moment to the dialysis equipment, so that the dialysis equipment can adjust the dialysis parameters of the target object according to the predicted dialysis parameters and / or predicted changes in dialysis parameters at the next moment.

[0177] Therefore, no human intervention is required, and dialysis parameters can be adjusted accurately.

[0178] In some embodiments, step 206 may be included after step 203.

[0179] Step 206: When the physiological state information of the target object reflects an abnormal situation, output alarm information.

[0180] Based on the historical data of the target object, abnormalities in the target object's physiological state information can be identified, and an alarm message can be output when an abnormality is detected. This alarm message can be an alarm tone, text message, or voice message, and this application does not limit the specific format of the alarm message.

[0181] Furthermore, one possible implementation of step 206 is: obtaining the average value of the indicator corresponding to the target object and a preset offset threshold. Based on the target object's physiological state information, the average value of the indicator, and the preset offset threshold, determining the target object's physiological state. When the physiological state indicates an abnormal situation, outputting an alarm message.

[0182] Optionally, due to individual differences, the mean values ​​of physiological state information differ among individuals. Therefore, the mean value of this indicator and the preset offset threshold can be set based on the target object's historical data and / or the application scenario. This historical data refers to the target object's past physiological state information.

[0183] Furthermore, since physiological state information contains multiple types of information, the blood flow status of the target object can be determined based on the target object's physiological state information, the mean value of the indicators, and the preset offset threshold, in the following way:

[0184] The physiological state information from multiple types is weighted and summed to obtain a comprehensive physiological state information value. If the absolute value of the difference between the comprehensive physiological state information value and the mean value of the indicator is greater than a preset offset threshold, the current situation is determined to be abnormal, and an alarm message can be output. If the absolute value of the difference between the comprehensive physiological state information value and the mean value of the indicator is not greater than the preset offset threshold, the current situation is determined to be normal.

[0185] Alternatively, it can be expressed by the following formula (11):

[0186]

[0187] Where Δd(t) represents the difference between the comprehensive value of physiological state information and the mean value of the index, δ is the preset offset threshold, and Alarm(t) is the real-time alarm indicator signal. Alarm(t) = 1 indicates that an alarm message is output. The real-time alarm indicator signal is fed back to the inspected personnel and operators in real time through the human-machine interface, reminding them to pay attention to the physiological state of the target object.

[0188] In some embodiments, since the objects corresponding to the samples in the training set used to train the signal processing model vary, the training set contains multiple sample data and their respective labels. Typically, there are relatively more labels with intermediate values. This means that the signal processing model is more accurate when evaluating data of objects with intermediate values, but may be less accurate when evaluating data of objects that are not in the intermediate values. Since the adjustment equipment for hemodialysis parameters needs to be applied to different objects, the loss value during the training process of the signal processing model is determined based on the label distribution smoothing method to address the problem of imbalanced data distribution in the training samples.

[0189] Optionally, a Label Distribution Smoothing (LDS) strategy can be employed to smooth the label distribution and introduce sample weights, thereby improving overall regression performance. LDS first discretizes continuous labels, dividing the label range into several equally spaced intervals, called bins. This maps continuous labels to discrete numbers b∈{0,1,…,B-1}, where B is the total number of intervals, indicating that the label distribution in the regression task can be divided into B intervals. Assuming y... mi and y max The minimum and maximum values ​​of the labels in the training set are respectively, and the mapping of the bin number is as follows: the bin number b(y) corresponding to the y-th sample is given by the following formula (12):

[0190]

[0191] Calculate the frequency of each label in each bin for all samples. The empirical label distribution is obtained as follows (13):

[0192]

[0193] Where δ(·) represents the indicator function, N is the total number of samples in the training set, and b i This represents the bin number corresponding to the i-th bin and the y-th sample.

[0194] To alleviate the discontinuity of the label distribution, LDS uses a kernel density estimation method to perform a one-dimensional convolution operation on the original distribution, that is, to use a smoothing kernel function k (Gaussian kernel) to blur the label distribution and generate a smoothed label distribution.

[0195] Optional, smoothed label distribution It can be expressed by the following formula (14):

[0196]

[0197] in, Let k(j) be the convolution radius, and k(j) be the j-th value of the kernel function, satisfying the normalization condition. It represents the frequency of the label of all samples in the interval with the number bj.

[0198] Then, for each sample, b is calculated based on the interval corresponding to its label. i The weight of each sample's label is calculated as the reciprocal of the smoothed probability of that interval. The weight ω can be obtained using the following formula (15). i :

[0199]

[0200] Here, ε is a small constant to prevent division by zero.

[0201] LDS is incorporated into the training process by weighting the loss for each sample. The weighted mean squared error loss function L is constructed using the following formula (16). LDS :

[0202]

[0203] Where, f(x) i ) represents the model for sample x i The prediction, y i It corresponds to its tag value.

[0204] In some embodiments, the dialysis parameter prediction model is based on a Hidden Markov Model (HMM) and a Convolutional Neural Network (U-Net) structure.

[0205] The physiological state information x(t) at the current time t output by the signal processing model can be in the following format:

[0206] x(t)=[PS t ED t VTI t ,EF t PWV t SBP t DBP t ]

[0207] Among them, PS t ED t and VTI t These are the maximum blood flow velocity, minimum blood flow velocity, and blood flow velocity-time integral in the carotid artery, EF. t It is an ejection fraction index, which can be used to derive parameters of cardiac contractility, PWV. t It is the speed of pulse wave propagation, SBP t DBP t These are systolic blood pressure and diastolic blood pressure, respectively.

[0208] Obtain a dialysis history data sequence, which may include the dialysis parameter sequence of the target object over the last T time periods. Furthermore, the dialysis history data sequence may also include the dialysis parameter sequence from previous dialysis procedures of the target object.

[0209] The dialysis parameters of the target object over the last T time periods can be expressed in the following format:

[0210]

[0211] Among them, T i Q is the temperature of the dialysate. i U is the dialysate flow rate. i To achieve the target water removal rate, Na i This represents the sodium ion concentration in the dialysate.

[0212] The physiological state information x(t) at the current time t and the dialysis historical data sequence are both inputs to the dialysis parameter prediction model.

[0213] The following section describes how predictions are made using Hidden Markov Models and the U-Net structure.

[0214] Hidden Markov Models (HMMs) can be applied in the biomedical field. Therefore, physiological state modeling of hemodynamic and cardiac function changes in time series is performed based on HMMs. Each hidden state in the HMM represents a potential "circulatory-cardiac function state combination," which can include a stable state, a sympathetic activation state, and a volume overload state. HMMs can learn state transition probabilities from historical dialysis data, outputting the predicted probability distribution of the current state and each state at the next time step. This enables dynamic identification and evolution prediction of physiological states during dialysis, providing a state context basis for parameter prediction.

[0215] In the modeling of potential physiological states, the hidden state space is defined as follows: S = {s1, s2, ..., s} N}. Among them, s n This represents the nth hidden state.

[0216] Hidden Markov Models (HMMs) can obtain the probability distribution of the current first-cycle cardiac function state and the predicted next-time cardiac function states. The circulatory cardiac function states are also referred to as hidden states or states. The HMM includes the following elements:

[0217] State transition matrix A = [a ij ], where: a ij =P(s) t =s j |s t-1 =s i ), a ij Indicates from hidden state s i Transition to hidden state s j The probability of.

[0218] The observation probability model, where the states follow a standard normal distribution, can be expressed as: x t It is the observation vector at time t. It is state s t The mean vector of the following observations, ∑s t It is state s t The covariance matrix of the observed values.

[0219] The initial probability state distribution can be expressed as: π i =P(s1=s i ).

[0220] After being trained using the Baum-Welch algorithm, the Hidden Markov Model can realize a given current observation sequence x. t-T:t At that time, the state label of the maximum a posteriori estimate of the current state. It is the state with the highest probability in the next moment, which can be obtained by the following formula (17):

[0221]

[0222] Next, a multi-channel time-series U-Net is used as a regressor to calculate the current observed signal x. t and status labels The predicted offset of the dialysis parameters at the next moment is shown in the following formula (18):

[0223]

[0224] Where, Δψ t+1 The offset component for each dialysis parameter: Δψ t+1 =[ΔT t+1 ,ΔQ t+1 ,ΔU t+1, ΔNa t+1

[0225] The U-Net input formats the micro motion signal spectrum. Among them, the micro motion signal spectrum can be a two-dimensional matrix of time × channel. Among them, each physiological signal is a channel. For example, the physiological signals can include: blood flow velocity (BFV), pulse wave velocity (PWV), mean arterial pressure estimation (MAP), etc. And introduce the current hidden state As an additional conditional vector, it enters the network coding path through the conditional embedding mechanism.

[0226] When training the dialysis parameter prediction model, historical annotation data (historical dialysis parameter sequence) can be used, and state-aware weights are added to construct the supervised learning objective function as the following formula (19):

[0227]

[0228] Among them, N is the number of the second samples, s i is the physiological state to which the i-th second sample belongs, is the weight of the state s i , is the predicted change in dialysis parameters output by the dialysis parameter prediction model, is the label, usually the actual recorded or artificially set ideal change in dialysis parameters.

[0229] Finally, the dialysis parameter offset is obtained and used to update the dialysis strategy in the next control cycle:

[0230] ψ t+1 = ψ t + Δψ t+1

[0231] Among them, ψ t is the dialysis parameter at the current time t, Δψ t+1 is the change in dialysis parameters from time t to time t + 1, and ψ t+1 is the dialysis parameter at the next moment of the current time t.

[0232] The HMM can achieve interpretable and hierarchical physiological state modeling. The U-Net network can capture complex spatial features and trend information in the time-series micro motion signals and be used in the real-time linkage control unit during dialysis to improve the individualization level of dialysis treatment. The dialysis device can automatically or semi-automatically adjust the parameter settings according to this control signal to form a data-driven intelligent closed-loop control system.

[0233] In some embodiments, the present application provides a millimeter-wave radar, which includes:

[0234] ​Antenna array for transmitting millimeter waves toward the blood vessels and heart of a target object and receiving the echo signals of the millimeter waves;

[0235] The processor is configured to, based on the echo signal, acquire a first signal corresponding to a first point reflecting the vibration of the blood vessel, and a second signal corresponding to a target region reflecting the vibration of the heart; input the first signal, the second signal, and the physiological information of the target object into a signal processing model, so that the signal processing model determines the physiological state information of the target object at the current moment, the physiological state information including hemodynamic characteristics and cardiac function characteristics; input the physiological state information, the dialysis parameters of the target object at the current moment, the historical dialysis parameters of the target object, and the corresponding historical physiological state information into a dialysis parameter prediction model, so that the dialysis parameter prediction model outputs the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

[0236] In an optional embodiment, the blood vessel includes a carotid artery; the processor, based on the echo signal, acquires a first signal corresponding to a point reflecting the vibration of the blood vessel, and a second signal corresponding to a target region reflecting the vibration of the heart tissue, including:

[0237] The processor processes the echo signal to obtain an intermediate frequency (IF) signal corresponding to the echo signal; based on the IF signal, it determines the spatial orientation spectrum corresponding to the target object; based on the relative positional relationship between the carotid artery and the heart, it divides the spatial orientation spectrum into a first spatial orientation spectrum corresponding to a first region and a second spatial orientation spectrum corresponding to a second region; the first region contains the carotid artery, and the second region contains the heart; based on the first spatial orientation spectrum, it determines the first point with the strongest intensity in the first spatial orientation spectrum and determines the location information of the first point; based on the location information of the first point, it extracts the first signal corresponding to the first point from the IF signal; based on the second spatial orientation spectrum, it determines the second point with the strongest intensity in the second spatial orientation spectrum and determines the location information of the second point; based on the location information of the second point, it determines the location information of the target region reflecting the vibration of the heart; based on the location information of the target region, it extracts the second signal corresponding to the target region from the IF signal.

[0238] In an optional embodiment, the processor determines the spatial orientation spectrum corresponding to the target object based on the intermediate frequency signal, including:

[0239] The processor determines the distance range of each distance cell based on the intermediate frequency signal; determines the distance cell to which the target object belongs based on the preset distance range between the target object and the target object; and determines the spatial orientation spectrum within the distance cell to which the target object belongs as the spatial orientation spectrum corresponding to the target object based on beamforming.

[0240] The processor determines the second point with the strongest intensity in the second spatial orientation spectrum based on the second spatial orientation spectrum, including:

[0241] The processor sums the power of each coordinate position in the second spatial orientation spectrum within the heart rate frequency band to obtain the signal energy spectrum within the heart rate frequency band; and obtains the second point with the greatest intensity in the signal energy spectrum.

[0242] In an optional embodiment, the hemodynamic features include blood flow velocity features, which include maximum blood flow velocity, minimum blood flow velocity, and blood flow velocity time-velocity integral.

[0243] In an optional embodiment, the cardiac function characteristics include: ejection fraction, pulse wave velocity, systolic blood pressure, and diastolic blood pressure;

[0244] In an optional embodiment, the dialysis parameters include: dialysate temperature, dialysate flow rate, target dehydration amount, and dialysate sodium ion concentration.

[0245] In an optional embodiment, the dialysis parameter prediction model is based on a Markov model and is obtained after training and convergence. The dialysis parameter prediction model is used to determine the probability distribution of the current first circulatory cardiac function state and the predicted circulatory cardiac function state at the next moment, and to obtain the maximum a posteriori estimated circulatory cardiac function state. Based on the current first circulatory cardiac function state and the maximum a posteriori estimated circulatory cardiac function state, the predicted change in dialysis parameters at the next moment is obtained.

[0246] In an optional embodiment, before the processor inputs the first signal, the second signal, and the physiological information of the target object into the signal processing model, it further includes:

[0247] The processor performs a first preprocessing on the first signal to obtain a preprocessed first signal; the first preprocessing is used to filter out noise signals caused by factors other than vascular vibration in the target signal; the processor performs a second preprocessing on the second signal to obtain a preprocessed second signal; the second preprocessing is used to filter out noise signals caused by factors other than heartbeat in the target signal.

[0248] The processor inputs the first signal, the second signal, and the physiological information of the target object into the signal processing model, which includes:

[0249] The processor inputs the preprocessed first signal, the preprocessed second signal, and the physiological information of the target object into the signal processing model.

[0250] In an optional embodiment, the first preprocessing includes at least one of a first acceleration differential filtering process, a first outlier processing process, and a first root mean square normalization process; the first acceleration differential filtering process is used to filter out noise in the first signal caused by respiratory displacement and / or human body movement; the outlier processing process is used to remove outliers from the first signal; and the root mean square normalization process is used to normalize the first signal according to a preset amplitude ratio.

[0251] If the first preprocessing includes first acceleration differential filtering, first outlier processing, and first root mean square normalization, then the processor performs the first preprocessing on the first signal to obtain a preprocessed first signal, including:

[0252] The processor obtains the filtered first signal according to the following formula:

[0253]

[0254] Among them, Diff_p i p is the first signal after filtering. i Let be the spatial spectrum at time i, and T be the time interval of the intermediate frequency signal;

[0255] The filtered first signal is divided according to a preset time window;

[0256] If the maximum absolute amplitude of the target time window is greater than the preset energy threshold, the first signal corresponding to the target time window is removed from the first signal after filtering to obtain the first signal after outlier processing.

[0257] The first signal after outlier processing is subjected to root mean square normalization to obtain the preprocessed first signal.

[0258] The second preprocessing includes at least one of a second acceleration differential filtering process, a second outlier processing process, and a second root mean square normalization process; the second acceleration differential filtering process is used to filter out noise caused by respiratory displacement and / or human body movement in the second signal; the outlier processing process is used to remove outliers from the second signal; the second root mean square normalization process is used to normalize the second signal according to a preset amplitude ratio;

[0259] If the second preprocessing includes second acceleration differential filtering, second outlier processing, and second root mean square normalization, then the second preprocessing of the second signal to obtain the preprocessed second signal includes:

[0260] The filtered second signal is obtained according to the following formula:

[0261]

[0262] Among them, Diff_p i p is the second signal after filtering. i Let be the spatial spectrum at time i, and T be the time interval of the intermediate frequency signal;

[0263] The filtered second signal is divided according to a preset time window;

[0264] If the maximum absolute amplitude of the target time window is greater than the preset energy threshold, the second signal corresponding to the target time window is removed from the filtered second signal to obtain the outlier-processed second signal.

[0265] The second signal after outlier processing is subjected to root mean square normalization to obtain the preprocessed second signal.

[0266] In an optional embodiment, the millimeter-wave radar further includes:

[0267] The sending module is used to send the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment to the dialysis device, so that the dialysis device adjusts the dialysis parameters of the target object according to the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

[0268] The millimeter-wave radar provided in this embodiment has a similar implementation principle and beneficial effects to the above-described method embodiments, and will not be repeated here.

[0269] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 3 As shown, in practice, this electronic device includes a memory 21 and a processor 22.

[0270] Memory 21 is used to store computer programs and can be configured to store various other data to support operation on the electronic device. Examples of this data include instructions for any application or method used to operate on the electronic device, data structures, contact data, phone book data, messages, pictures, videos, etc.

[0271] The processor 22, coupled to the memory 21, is used to execute the computer program in the memory 21 to implement the method for adjusting hemodialysis parameters provided in the foregoing embodiments.

[0272] Furthermore, such as Figure 3 As shown, the electronic device also includes other components such as a communication component 23, a display 24, a power supply component 25, and an audio component 26. Figure 3 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 3 The components shown are as follows. The electronic device in this embodiment can be a terminal device such as a desktop computer, laptop computer, smartphone, or IoT device, or a server device such as a conventional server, cloud server, or server array.

[0273] The aforementioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0274] The aforementioned communication component is configured to facilitate wired or wireless communication between the device containing the communication component and other devices. The device containing the communication component can access wireless networks based on communication standards, such as 2G, 3G, 4G / LTE, 5G, or combinations thereof. In one exemplary embodiment, the communication component receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel.

[0275] The aforementioned display includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a Touch Panel, the screen can be implemented as a touchscreen to receive input signals from the user. The Touch Panel includes one or more touch sensors to sense touches, swipes, and gestures on the Touch Panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation.

[0276] The aforementioned power supply components provide power to various components within the device in which they reside. These power supply components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to the device in which they reside.

[0277] The aforementioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, or voice recognition mode. The received audio signals can be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals.

[0278] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the steps in the above-described method embodiments. The computer-readable storage medium includes volatile or non-volatile components, or a combination thereof, and can be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), flash memory or other memory technologies, CD-ROM, Digital Video Disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium.

[0279] Accordingly, this application also provides a computer program product, which includes a computer program or instructions that, when executed by a processor, cause the processor to implement the steps in the above method embodiments. It should be understood that each step or combination of steps in the above method flow can be implemented by the computer program or instructions. Furthermore, these computer programs or instructions can be applied to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device, enabling the processor of the general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to function as an apparatus for implementing the corresponding functions in the above method embodiments.

[0280] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for adjusting hemodialysis parameters, characterized in that, The method includes: Millimeter waves are emitted toward the blood vessels and heart of the target object, and the echo signals of the millimeter waves are received; Based on the echo signal, a first signal corresponding to a first point reflecting the vibration of the blood vessel and a second signal corresponding to a target area reflecting the vibration of the heart are obtained from the echo signal. The first signal, the second signal, and the physiological information of the target object are input into a signal processing model so that the signal processing model can determine the physiological state information of the target object at the current moment, the physiological state information including hemodynamic characteristics and cardiac function characteristics; The physiological state information, the current dialysis parameters of the target object, the historical dialysis parameters of the target object, and the corresponding historical physiological state information are input into the dialysis parameter prediction model so that the dialysis parameter prediction model outputs the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

2. The method according to claim 1, characterized in that, The blood vessel includes the carotid artery; the step of obtaining a first signal corresponding to a first point reflecting the vibration of the blood vessel and a second signal corresponding to a target region reflecting the vibration of the heart tissue from the echo signal includes: The echo signal is processed to obtain the intermediate frequency signal corresponding to the echo signal; Based on the intermediate frequency signal, determine the spatial orientation spectrum corresponding to the target object; Based on the relative positional relationship between the carotid artery and the heart, the spatial orientation spectrum is divided into a first spatial orientation spectrum corresponding to a first region and a second spatial orientation spectrum corresponding to a second region; the first region contains the carotid artery and the second region contains the heart; Based on the first spatial orientation spectrum, determine the first point with the strongest intensity in the first spatial orientation spectrum, and determine the location information of the first point; Based on the location information of the first point, extract the first signal corresponding to the first point from the intermediate frequency signal; Based on the second spatial orientation spectrum, determine the second point with the strongest intensity in the second spatial orientation spectrum, and determine the location information of the second point; Based on the location information of the second point, determine the location information of the target area reflecting the vibration of the heart; Based on the location information of the target area, the second signal corresponding to the target area is extracted from the intermediate frequency signal.

3. The method according to claim 2, characterized in that, The step of determining the spatial orientation spectrum corresponding to the target object based on the intermediate frequency signal includes: Based on the intermediate frequency signal, the distance range of each distance chamber is determined; Determine the distance bin to which the target object belongs based on the preset distance range between the target object and the target object; Based on beamforming, the spatial orientation spectrum within the range cell to which the target object belongs is determined as the spatial orientation spectrum corresponding to the target object; The step of determining the second point with the strongest intensity in the second spatial orientation spectrum based on the second spatial orientation spectrum includes: The power at each coordinate position in the second spatial orientation spectrum within the heart rate frequency band is summed to obtain the signal energy spectrum within the heart rate frequency band. Obtain the second point with the highest intensity in the energy spectrum of the signal.

4. The method according to any one of claims 1-3, characterized in that, The hemodynamic features include blood flow velocity features, which include maximum blood flow velocity, minimum blood flow velocity, and blood flow velocity time-velocity integral. The cardiac function characteristics include: ejection fraction, pulse wave velocity, systolic blood pressure, and diastolic blood pressure; The dialysis parameters include: dialysate temperature, dialysate flow rate, target dehydration amount, and dialysate sodium ion concentration.

5. The method according to any one of claims 1-3, characterized in that, The dialysis parameter prediction model is based on a Markov model and is obtained after training and convergence. The dialysis parameter prediction model is used to determine the probability distribution of the current first circadian cardiac function state and the predicted next circadian cardiac function state, and to obtain the maximum a posteriori estimated circadian cardiac function state. Based on the current first circadian cardiac function state and the maximum a posteriori estimated circadian cardiac function state, the predicted change in dialysis parameters at the next circadian time is obtained.

6. The method according to any one of claims 1-3, characterized in that, Before inputting the first signal, the second signal, and the physiological information of the target object into the signal processing model, the method further includes: The first signal is subjected to a first preprocessing to obtain a preprocessed first signal; the first preprocessing is used to filter out noise signals caused by factors other than vascular vibration in the target signal. The second signal is subjected to a second preprocessing to obtain a preprocessed second signal; the second preprocessing is used to filter out noise signals caused by factors other than heartbeat in the target signal. The step of inputting the first signal, the second signal, and the physiological information of the target object into the signal processing model includes: The preprocessed first signal, the preprocessed second signal, and the physiological information of the target object are input into the signal processing model.

7. The method according to claim 6, characterized in that, The first preprocessing includes at least one of the following: first acceleration differential filtering, first outlier processing, and first root mean square normalization; the first acceleration differential filtering is used to filter out noise in the first signal caused by respiratory displacement and / or human body movement. The outlier processing is used to remove outliers from the first signal; the root mean square normalization processing is used to normalize the first signal according to a preset amplitude ratio; If the first preprocessing includes first acceleration differential filtering, first outlier processing, and first root mean square normalization, then the first preprocessing of the first signal to obtain the preprocessed first signal includes: The first signal after filtering is obtained according to the following formula: Among them, Diff_p i p is the first signal after filtering. i Let be the spatial spectrum at time i, and T be the time interval of the intermediate frequency signal; The filtered first signal is divided according to a preset time window; If the maximum absolute amplitude of the target time window is greater than the preset energy threshold, the first signal corresponding to the target time window is removed from the first signal after filtering to obtain the first signal after outlier processing. The first signal after outlier processing is subjected to root mean square normalization to obtain the preprocessed first signal. The second preprocessing includes at least one of a second acceleration differential filtering process, a second outlier processing process, and a second root mean square normalization process; the second acceleration differential filtering process is used to filter out noise caused by respiratory displacement and / or human body movement in the second signal; the outlier processing process is used to remove outliers from the second signal; the second root mean square normalization process is used to normalize the second signal according to a preset amplitude ratio; If the second preprocessing includes second acceleration differential filtering, second outlier processing, and second root mean square normalization, then the second preprocessing of the second signal to obtain the preprocessed second signal includes: The filtered second signal is obtained according to the following formula: Among them, Diff_p i p is the second signal after filtering. i Let be the spatial spectrum at time i, and T be the time interval of the intermediate frequency signal; The filtered second signal is divided according to a preset time window; If the maximum absolute amplitude of the target time window is greater than the preset energy threshold, the second signal corresponding to the target time window is removed from the filtered second signal to obtain the outlier-processed second signal. The second signal after outlier processing is subjected to root mean square normalization to obtain the preprocessed second signal.

8. The method according to any one of claims 1-3, characterized in that, The method further includes: The dialysis device sends the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment to the dialysis device, so that the dialysis device adjusts the dialysis parameters of the target object according to the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

9. A millimeter-wave radar, characterized in that, The millimeter-wave radar includes: Antenna array for transmitting millimeter waves toward the blood vessels and heart of a target object and receiving the echo signals of the millimeter waves; The processor is configured to, based on the echo signal, acquire a first signal corresponding to a first point reflecting the vibration of the blood vessel, and a second signal corresponding to a target region reflecting the vibration of the heart; The first signal, the second signal, and the physiological information of the target object are input into a signal processing model so that the signal processing model can determine the physiological state information of the target object at the current moment, the physiological state information including hemodynamic characteristics and cardiac function characteristics; The physiological state information, the current dialysis parameters of the target object, the historical dialysis parameters of the target object, and the corresponding historical physiological state information are input into the dialysis parameter prediction model so that the dialysis parameter prediction model outputs the predicted dialysis parameters and / or the predicted changes in dialysis parameters at the next moment.

10. A non-transitory machine-readable storage medium, characterized in that, The non-transitory machine-readable storage medium stores executable code that, when executed by a processor of an electronic device, causes the processor to perform the method for adjusting hemodialysis parameters as described in any one of claims 1 to 8.

Images