Magnetic suspension cushioning road condition sensing and active suspension method for ambulance

Abstract

The invention provides a magnetic suspension cushioning road condition sensing and active suspension method for an ambulance, and the method comprises the steps: dynamically calculating the current intensity control quantity of each electromagnetic coil according to the injured part of a patient, the injury degree information and the conditions of other uninjured parts of the body through a fuzzy control algorithm, exerting small supporting force on the seriously injured part, and carrying out the dynamic suspension of the patient. Normal supporting force is maintained for uninjured parts; the stretcher position of the injured part is recognized, a directionally focused ultrasonic standing wave field is generated under the sensitive injured part through a phased array ultrasonic transducer array, miniature suspension points are formed on the surface of the body of a patient according to the acoustic suspension principle, local stress is further relieved, and secondary injury to the fracture part is avoided.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a magnetic suspension shock-absorbing road condition perception and active shock-absorbing method for an ambulance. Background Art

[0002] When designing the smart stretcher system on the ambulance, a key technical challenge is how to accurately control the current intensity distribution of the electromagnetic coil under the stretcher based on the real-time analysis of the patient's physiological data, so as to dynamically adjust the repulsive force distribution between the magnetic rings in the suspension device and provide differentiated support strength for different parts of the patient's body. This requires the establishment of a complex mathematical model to map the patient's physiological parameters to the spatial distribution function of the coil current intensity. The model must fully consider factors such as the anatomical differences in various parts of the human body, the uneven distribution of body weight, and different types and degrees of injury. In addition, there are many challenges in integrating ultrasonic standing wave technology into the suspension system. In ultrasonic standing wave technology, the emission and receiving ends of ultrasonic waves form a collision and then form a standing wave point. Under the repeated standing waves, a balance point is formed, which can make the object levitate. First, the optimal ultrasonic frequency and phase distribution must be found to form a stable acoustic suspension point around the injury while avoiding interference with other areas. Secondly, the propagation and attenuation characteristics of ultrasonic waves in human tissue are complex and changeable, and adaptive algorithms need to be developed to optimize the sound field distribution in real time. Furthermore, the coupling effect of electromagnetic suspension and ultrasonic suspension may lead to unpredictable dynamic behavior, and a lot of theoretical analysis and experimental verification must be carried out to find the best collaborative working mode of the two mechanisms. Finally, the power consumption and heat dissipation of the entire system cannot be ignored. It is necessary to optimize the design while ensuring the suspension performance and taking into account the energy efficiency and thermal management. Summary of the invention

[0003] The present invention provides a magnetic suspension shock-absorbing road condition perception and active shock-avoidance method for an ambulance, which mainly includes: Identify the patient's injured part and injury degree information, establish a mapping table between injury and current intensity in advance, and adjust the current intensity of the electromagnetic coils in different areas under the stretcher according to the electromagnetic coil current intensity control parameters corresponding to the mapping table between injury and current intensity, so as to change the local magnetic field distribution and affect the repulsive force between the magnetic rings in the neodymium iron suspension device; The fuzzy control algorithm is used to dynamically calculate the current intensity control amount of each electromagnetic coil according to the patient's injured part, injury degree information and the condition of other uninjured parts of the body, so as to apply a small support force to the seriously injured part and maintain normal support force for the uninjured part. Identify the stretcher position of the injured part, and generate a directional focused ultrasonic standing wave field through the phased array ultrasonic transducer array just below the sensitive injured part. Use the principle of acoustic levitation to form a micro-suspension point on the patient's body surface, further reducing local stress and avoiding secondary injury to the fracture site. The pressure sensor array detects the contact pressure distribution between the patient's body parts and the stretcher in real time, feeds the pressure data to the ultrasonic standing wave control system, and dynamically adjusts the focus position and sound intensity of the ultrasonic standing wave field, so that the levitation point is always aimed at the most seriously injured area, and changes the size of the acoustic levitation force as needed; The vibration of the stretcher is monitored in real time. The vibration frequency and amplitude data are obtained through the acceleration sensor. The data is compared with the preset safety threshold of the vibration frequency and amplitude data. If the threshold is exceeded, an alarm is triggered. At the same time, the vibration data is input into the electromagnetic suspension and ultrasonic suspension control system. The vibration is actively suppressed by adjusting the electromagnetic coil current and the ultrasonic standing wave sound intensity. Comprehensively analyze the changing trends of the patient's physiological data, changes in pressure distribution, and stretcher vibration to predict the progression of the patient's condition. When the prediction results show that the condition may worsen, adjust the electromagnetic suspension and ultrasonic suspension strategies in advance, increase the protection of parts that are vulnerable to secondary injuries, and issue early warnings to medical staff to provide a reference for subsequent treatment decisions.

[0004] The technical solution provided by the embodiment of the present invention may have the following beneficial effects: The present invention discloses a magnetic levitation shock-absorbing road condition perception and active shock-avoidance method for an ambulance. The method dynamically adjusts the distribution of the stretcher support force by identifying the patient's injured part and degree, combining electromagnetic suspension and ultrasonic suspension technology. Specifically, the present invention adopts a fuzzy control algorithm to control the current intensity of the electromagnetic coil according to the injury information, and changes the local magnetic field distribution; at the same time, a directionally focused ultrasonic standing wave field is generated below the injured part to form a micro-suspension point. The present invention also monitors the contact pressure distribution in real time through a pressure sensor array, and adjusts the ultrasonic standing wave field through feedback; and uses an acceleration sensor to monitor vibration and actively suppress excessive vibration. In addition, the present invention comprehensively analyzes physiological data, pressure distribution and vibration conditions, predicts the progression of the disease and adjusts the protection strategy in a timely manner. The system can significantly reduce the pressure on the injured part, prevent secondary injuries, and provide safety protection for the transportation of critically ill patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1 A flow chart of a magnetic suspension shock-absorbing road condition perception and active shock-avoidance method for an ambulance of the present invention; Figure 2 A schematic diagram of a magnetic suspension shock-absorbing road condition perception and active shock-avoidance method for an ambulance of the present invention; Figure 3It is another schematic diagram of the magnetic suspension shock-absorbing road condition perception and active shock-avoidance method of an ambulance of the present invention. DETAILED DESCRIPTION

[0006] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0007] like Figure 1-3 In this embodiment, a method for magnetic suspension shock-absorbing road condition perception and active shock-avoidance of an ambulance may specifically include: S101. Identify the patient's injured part and injury severity information, establish a mapping table between injury and current intensity in advance, and adjust the current intensity of the electromagnetic coils in different areas under the stretcher according to the electromagnetic coil current intensity control parameters corresponding to the mapping table between injury and current intensity to change the local magnetic field distribution to affect the repulsive force between the magnetic rings in the neodymium iron suspension device.

[0008] Acquire a patient image, use the YOLO algorithm to locate the injured part of the patient, and classify and evaluate the severity of the injury based on the positioning result; determine the current intensity control parameters required by the electromagnetic coil based on the injury severity assessment result and the pre-established injury and current intensity lookup table mapping; calculate the specific area coordinates under the stretcher corresponding to the injured part based on the position information of the patient's injured part and the size of the stretcher, and distribute the current intensity to the electromagnetic coils around the area; perform finite element analysis on the current intensity distribution result, solve the Maxwell equations, and obtain the magnetic flux density in different areas under the stretcher; use the magnetic flux density distribution to perform multi-body dynamics simulation on the suspension system, determine the equilibrium position between the magnetic rings through iterative calculation, and obtain the suspension height distribution of each point on the stretcher.

[0009] Specifically, the patient image is obtained and the YOLO algorithm is used to locate the injured part. The severity of the injury is classified and evaluated through ResNet. The current intensity control parameters required for the electromagnetic coils in different areas are determined based on the pre-established table mapping of the injury and the current intensity. According to the location information of the patient's injured part and the size of the stretcher, the coordinates of the specific area under the stretcher corresponding to the injured part are calculated, and the current intensity of the electromagnetic coils around the area is distributed using the bicubic interpolation algorithm to generate a continuous and smooth current intensity distribution surface. Finite element analysis is performed using COMSOL Multiphysics software to establish a geometric model, set material properties, define boundary conditions, generate a grid, solve Maxwell's equations, post-process the results, calculate the magnetic field distribution under the stretcher, and obtain the magnetic flux density in different areas. The Biot-Savart law is used to calculate the magnitude and direction of the magnetic force acting on the neodymium iron magnetic ring, and the combined force of gravity and magnetism is considered to obtain the force condition of each magnetic ring. According to the force analysis results of the neodymium iron ring, Adams software was used to perform multi-body dynamics simulation on the suspension system, establish a suspension system model including the magnetic ring and the connection structure, set the mass and inertia parameters of each component, define constraints and forces, determine the equilibrium position between the magnetic rings through iterative calculation, and finally obtain the suspension height distribution of each point on the stretcher. When the YOLO algorithm locates the injured part, the input is a 1920x1080 pixel RGB image, the features are extracted through the convolution layer, and then the target frame coordinates are obtained through non-maximum suppression, such as x=500, y=300, width=200, height=150. The ResNet network receives a 224x224 pixel image block and outputs a probability distribution of injury severity in 5 categories, such as [0.1, 0.2, 0.5, 0.15, 0.05]. The table lookup mapping converts the severity into current intensity, for example, a moderate injury corresponds to a current of 2.5A. The size of the stretcher is 2000x800mm, and the coordinates of the injured part (500,300) are mapped to the stretcher coordinate system as (625,300). The bicubic interpolation algorithm uses 16 neighboring points to generate a 100x40 current intensity matrix. In COMSOL software, the relative magnetic permeability of air is set to 1, the relative magnetic permeability of the NdFeB magnetic ring is set to 1.05, and the boundary condition is magnetic insulation. The mesh is divided into tetrahedral elements, with a total number of about 500,000 elements. The three-dimensional magnetic field distribution is obtained by solving, such as the magnetic induction intensity of 0.5T at the point (100,200,50). The magnetic force is calculated using the Biot-Savart law, considering the vector sum of gravity 9.8N and magnetic force 2.3N. A suspension system model of 10 magnetic rings is established in Adams software, each with a mass of 0.5kg and a moment of inertia of a diagonal matrix [0.001,0.001,0.002]kg·m². Define the spring constraint between the magnetic rings, with a stiffness coefficient of 1000N / m. After 100 iterations, the suspension height distribution of each point of the stretcher in the stable state is obtained, such as the center point height is 150mm, and the four corner points height is 145mm.

[0010] S102. Using a fuzzy control algorithm, the current intensity control amount of each electromagnetic coil is dynamically calculated according to the patient's injured part, injury degree information and the condition of other uninjured parts of the body, so as to apply a small supporting force to the seriously injured part and maintain a normal supporting force for the uninjured part.

[0011] A whole-body image of the patient is acquired, the whole-body image of the patient is scanned, the position coordinates and the area size of the injured area are obtained, and the injured area is quantitatively evaluated to determine the numerical value of the injury degree; a Mamdani fuzzy inference system is established according to the numerical value of the injury degree and the area size of the injured area, input variables and output variables are set, the input variables are fuzzified, and a fuzzy set of the output variables is obtained; a relationship model between current intensity and magnetic field intensity is established according to the fuzzy set of the output variables, the magnetic field intensity is calculated and the magnetic force is obtained to determine the size of the supporting force; the current size of the electromagnetic coils in different areas under the stretcher is adjusted in real time through a PID controller, the local magnetic field distribution is changed, the repulsive force between the magnetic rings in the neodymium iron suspension device is affected, and a supporting force is applied to the injured area.

[0012] Specifically, the patient's injured part and injury degree information is obtained, and the patient's whole body is scanned through ResNet50 to identify the location coordinates and area size of the injured area, and the injury degree is quantitatively evaluated according to the AIS scale to obtain the severity value of the injured part. A Mamdani fuzzy inference system is established, and the input variables are set as "injury degree" and "injury area", and the output variable is "current intensity". Fuzzy rules are formulated through data mining methods, such as "if the injury degree is severe and the injury area is large, the current intensity is low". The fuzzy inference engine is used to fuzzify the input variables, calculate the membership values ​​of each input variable under different membership functions, use the minimum and maximum synthesis method for fuzzy inference, obtain the fuzzy set of output variables, and then use the center of gravity method for defuzzification to obtain the precise current intensity control quantity of each electromagnetic coil. According to the current intensity control quantity, a relationship model between current intensity and magnetic field intensity is established, the magnetic field intensity is calculated and the magnetic force is obtained, and finally the supporting force is obtained. The PID controller is used to adjust the current of the electromagnetic coils in different areas under the stretcher in real time, changing the local magnetic field distribution to affect the repulsive force between the magnetic rings in the neodymium iron suspension device, applying a smaller support force to the severely injured part and maintaining a normal support force for the uninjured part. ResNet50 receives RGB images of 1920x1080 pixels, extracts features through convolutional layers and pooling layers, and outputs the coordinates of the injured area such as (500,300,200,150) and an area of ​​30,000 square pixels. The AIS scale divides injuries into levels 1-6, corresponding to light, moderate, and severe levels, such as level 4 corresponding to severe injuries. In the Mamdani fuzzy inference system, three fuzzy sets are set for the degree of injury and the area of ​​injury, and five fuzzy sets are set for the current intensity. 25 fuzzy rules are formulated through 200 expert questionnaires and 10,000 historical data. After the input variables are fuzzified, the membership degree of the "severe" fuzzy set for the level 4 injury is 0.8. The activation intensity of the rules is calculated by the minimum and maximum synthesis method. For example, the activation intensity of the rule "If the injury is severe and the injured area is large, the current intensity is low" is 0.7. The center of gravity method is used to defuzzify and obtain the current intensity control quantity, such as 2.5A. The current and magnetic field model B=kI is established, where the proportional coefficient k is 0.01T / A, I is the current intensity, and the magnetic field intensity B is calculated to be 0.025T. The magnetic force calculation formula F=B^2A / (2*μ0), μ0 is the vacuum magnetic permeability, A is the magnetic pole area, and the support force F is 1.5N. The PID controller parameters are set to Kp=0.5, Ki=0.1, Kd=0.05, the sampling period is 0.1s, and the current is adjusted in real time to stabilize the support force at the target value. Through dynamic adjustment, the support force of severely injured parts is reduced to 60% of the normal value, and the uninjured parts maintain 100% support force.

[0013] The patient's injured part, injury degree and uninjured part are collected as input variables, and the input variables are fuzzified through the membership function to obtain the fuzzy control quantity; a fuzzy rule base is established, and the fuzzy reasoning mechanism is used to derive the current intensity control quantity of each electromagnetic coil, and the fuzzy control quantity is converted into an accurate current value. The output current intensity is inversely proportional to the severity of the patient's injury.

[0014] The position coordinates and area size of the injured part are determined according to the whole body scan image of the patient; the AIS scale evaluation value of the injured part is obtained, and the severity value of the injured part is determined by the AIS scale evaluation value; a Mamdani fuzzy inference system is established, and the Mamdani fuzzy inference system includes an injury degree input variable, an injury area input variable and an electromagnetic coil current intensity output variable; the injury degree input variable and the injury area input variable are converted into fuzzy sets using the membership function of the Mamdani fuzzy inference system; fuzzy inference is performed on the fuzzy set by the minimum and maximum synthesis method to obtain a fuzzy set of the electromagnetic coil current intensity output variable; the fuzzy set of the electromagnetic coil current intensity output variable is defuzzified by the centroid method; and the precise current intensity control quantity of each electromagnetic coil is determined by the defuzzification process.

[0015] Specifically, the ResNet50 network was used to scan the patient's whole body to obtain the position coordinates and area size of the injured part. The degree of injury was quantitatively evaluated according to the AIS scale to obtain the severity value of the injured part, and the status information of the uninjured part was recorded at the same time. The Mamdani fuzzy inference system was established, and the degree of injury and the injured area were set as input variables, and the current intensity of the electromagnetic coil was set as the output variable. Fuzzy rules such as "if the degree of injury is severe and the injured area is large, the current intensity is low" were constructed through data mining methods, and corresponding membership functions were defined for each variable. The input variables were converted into fuzzy sets using membership functions, and the membership values ​​of each input variable under different membership functions were calculated. The activated rules were fuzzy inferenced through the minimum and maximum synthesis method to obtain the fuzzy set of output variables. The center of gravity method is used to defuzzify the output fuzzy set, and the precise current intensity control quantity of each electromagnetic coil is calculated. The inverse relationship between the current intensity I and the severity of the injury is established through the exponential function I=Imax*e^(-k*S), where I is the current intensity, Imax is the maximum current value, k is the attenuation coefficient, and S is the severity of the injury, so as to apply a smaller support force to the severely injured part and maintain a normal support force for the uninjured part. The 1920x1080 pixel RGB image is received through the ResNet50 network, and the features are extracted through the convolution layer and the pooling layer, and the coordinates of the injured area such as (500,300,200,150) and the area of ​​30,000 square pixels are output. The AIS scale divides the injury into 1-6 levels, corresponding to the three levels of light, moderate and severe, such as level 4 corresponding to severe injury. In the Mamdani fuzzy inference system, the degree of injury and the area of ​​injury are set to three fuzzy sets, including light, medium, severe and small, medium, large, and the current intensity is set to five fuzzy sets, including very low, low, medium, high, and very high. 25 fuzzy rules were formulated through 200 expert questionnaires and 10,000 historical data. After the input variables were fuzzified, for example, the degree of membership of the "severe" fuzzy set for the degree of injury level 4 was 0.8, and the degree of membership of the "large" fuzzy set for the injured area of ​​30,000 square pixels was 0.7. The rule activation strength was calculated by the minimum and maximum synthesis method, such as the activation strength of the rule "If the degree of injury is severe and the injured area is large, the current intensity is low" is min(0.8,0.7)=0.7. The center of gravity method was used to defuzzify the current intensity control amount, such as 2.5A. In the exponential function, Imax was set to 5A, k was 0.5, and when S was 4 (severe injury), I was calculated to be about 1.65A, which achieved the reduction of the current intensity of the severely injured part. Through dynamic adjustment, the supporting force of severely injured parts is reduced to 33% of the normal value, while the uninjured parts maintain 100% supporting force.

[0016] S103. Identify the stretcher position of the injured part. Directly below the sensitive injured part, generate a directionally focused ultrasonic standing wave field through the phased array ultrasonic transducer array. Use the principle of acoustic levitation to form micro-suspension points on the patient's body surface to further reduce local stress and avoid secondary injury to the fracture site.

[0017] The YOLOv5 algorithm is used to analyze the patient image to identify the position and range information of the injured part; according to the position and range information of the injured part, it is mapped to the stretcher coordinate system to obtain the position data of the injured part on the stretcher; according to the position data of the injured part on the stretcher, a genetic algorithm is used to optimize the layout scheme of the phased array ultrasonic transducer array, and the control parameters of the transducer array are determined; based on the control parameters of the transducer array, a numerical simulation of the sound field is performed to obtain the sound pressure distribution; according to the sound pressure distribution, the phased array ultrasonic transducer array is driven by a sound field controller to generate a directionally focused ultrasonic standing wave field, and a micro-suspended point array is formed on the surface of the patient's body.

[0018] Specifically, the YOLOv5 algorithm is used to analyze the patient image, identify the specific location and range of the injured part, and map it to the stretcher coordinate system to obtain the precise location information of the injured part on the stretcher. According to the location, shape and size information of the injured part, the genetic algorithm is used to optimize the layout of the phased array ultrasonic transducer array, calculate the optimal installation position and orientation of each transducer, and generate the control parameters of the transducer array, including phase difference and amplitude. The sound field is numerically simulated by COMSOL Multiphysics software, the parameters of the transducer array and the position of the target suspension point are input, the Helmholtz equation is solved, the sound pressure distribution is obtained, and the optimal focusing parameters are determined. Among them, according to the transducer array parameters and the position of the target suspension point, the sound field numerical simulation model is established in COMSOL Multiphysics software. The physical parameters of the sound wave propagation medium, including density, sound speed, etc., are set in the model, and the boundary conditions of the model are defined. The finite element method is used to mesh the model and generate a computational grid. A sound source is added to the model, and the position and amplitude of the sound source are set according to the transducer array parameters. The sound source can be a point sound source, a surface sound source or a volume sound source. Solve the Helmholtz equation to obtain the numerical solution of the sound field. The Helmholtz equation describes the propagation law of sound waves in the medium. The calculation results are post-processed to obtain a sound pressure distribution cloud map. According to the sound pressure distribution, the focusing of the sound wave near the target suspension point position is judged. Through optimization algorithms, such as genetic algorithms or particle swarm algorithms, the transducer array parameters, such as the number of array elements, spacing, amplitude and phase, are adjusted to obtain the best focusing parameter combination. The phased array ultrasonic transducer array is driven by a sound field controller to generate a directionally focused ultrasonic standing wave field according to the calculated focusing parameters, forming a micro-suspended point array on the patient's body surface. The position and stability of the suspension point are monitored in real time by a pressure sensor, and the PID controller is used to adjust the sound field intensity and distribution to maintain a stable suspension effect. For the fracture site, a sound pressure safety threshold is set to limit the maximum sound pressure and avoid secondary injury. The YOLOv5 algorithm receives a 1920x1080 pixel RGB image, extracts features through a convolutional layer and a feature pyramid network, and outputs the bounding box coordinates of the injured part, x=500, y=300, width=200, height=150 and a confidence level of 0.95. The pixel coordinates are converted to the stretcher coordinate system to obtain the center point of the injured part (62.5cm, 37.5cm). The genetic algorithm takes maximizing the sound field intensity as the objective function, sets the population size to 100, iterates 500 times, and calculates the optimal position and orientation of 16 transducers. In the COMSOL software, the air density is set to 1.225kg / m³, the sound speed is 343m / s, and the frequency is 40kHz. The three-dimensional sound pressure distribution is solved, and the maximum sound pressure value is 1000Pa. The sound field controller drives the transducer array with an output power of 5W and a phase accuracy of 0.1°.The pressure sensor has a sampling frequency of 1kHz, a measurement range of 0-2000Pa, and a resolution of 1Pa. The PID controller parameters are set to Kp=0.5, Ki=0.1, Kd=0.05, and a sampling period of 1ms. The sound field intensity is adjusted in real time to control the position error of the suspension point within ±0.1mm. The sound pressure safety threshold of the fracture site is set to 800Pa. When it is detected that the threshold is exceeded, the controller automatically reduces the sound intensity of the area by 10%. Through dynamic adjustment, 20 micro suspension points are formed on the surface of the patient's body, each with a diameter of 2mm and a bearing capacity of 0.5g.

[0019] S104. The contact pressure distribution between various parts of the patient's body and the stretcher is detected in real time through the pressure sensor array, and the pressure data is fed back to the ultrasonic standing wave control system to dynamically adjust the focal position and sound intensity of the ultrasonic standing wave field so that the levitation point is always aligned with the most seriously injured area, and the size of the acoustic levitation force can be changed as needed.

[0020] The contact pressure data between various parts of the patient's body and the stretcher are obtained; according to the contact pressure data, the pressure distribution is analyzed to obtain the center coordinates and pressure values ​​of the pressure abnormality area and the most seriously injured part; after obtaining the center coordinates and pressure values ​​of the pressure abnormality area and the most seriously injured part, a pressure distribution thermodynamic map is generated according to the center coordinates and pressure values; through the pressure distribution thermodynamic map, the particle swarm optimization algorithm is used to calculate the optimal focal position and the required sound intensity of the ultrasonic standing wave field, and the optimal focal position and the required sound intensity are used to optimize the driving parameters of the ultrasonic transducer array; the driving parameters are transmitted to the ultrasonic standing wave control system, and real-time feedback control is realized through the PID controller to dynamically adjust the phase and amplitude of the transducer array; according to the real-time data of the pressure sensor and the pre-established mapping relationship between the pressure value and the sound intensity, the sound intensity is adjusted to realize the dynamic change of the acoustic levitation force.

[0021] Specifically, a high-precision pressure sensor array is arranged on the surface of the stretcher to collect the contact pressure data between various parts of the patient's body and the stretcher in real time. The pressure signal is converted into a digital signal through the data acquisition module, and the data is transmitted to the central processing unit. The convolutional neural network with the U-Net architecture is used to analyze the pressure distribution data, identify the pressure abnormality area and the most seriously injured part, calculate the center coordinates and pressure values ​​of these areas, and generate a pressure distribution heat map. According to the pressure distribution heat map, the particle swarm optimization algorithm is used to calculate the optimal focal position and required sound intensity of the ultrasonic standing wave field, and the sound pressure acoustic module of the COMSOL Multiphysics software is used to simulate the sound field distribution under different parameters to optimize the driving parameters of the ultrasonic transducer array. Real-time feedback control is achieved through the PID controller, and the optimized driving parameters are transmitted to the ultrasonic standing wave control system. The phase and amplitude of the transducer array are dynamically adjusted to accurately align the suspension point with the most seriously injured area. A mapping relationship between pressure value and sound intensity is established, and the sound intensity is adjusted according to the real-time data of the pressure sensor to achieve dynamic changes in the size of the acoustic levitation force. The high-precision pressure sensor array contains 1000 sensors with a sensitivity of 0.1N / cm² and a sampling frequency of 100Hz. The analog signal is converted into a digital signal through a 16-bit ADC. The U-Net network contains 4 layers of downsampling and 4 layers of upsampling, uses the ReLU activation function, inputs a 100x100 pressure matrix, and outputs a thermal map of the same size. The network identifies 3 pressure anomaly areas with center coordinates of (25,30), (60,45), and (80,70), corresponding to pressure values ​​of 2.5N / cm², 3.2N / cm², and 1.8N / cm². The particle swarm optimization algorithm sets 100 particles and iterates 500 times to calculate the optimal focus position (62,48) and the required sound intensity of 85dB. In the COMSOL simulation, the air density is set to 1.225kg / m³, the sound speed is 343m / s, and the frequency is 40kHz. The sound field distribution is calculated, and the maximum sound pressure value is 1000Pa. The PID controller parameters are set to Kp=0.5, Ki=0.1, Kd=0.05, and the sampling period is 10ms. The phase (accuracy 0.1°) and amplitude (range 0-5W) of the 16 transducers are adjusted in real time according to the pressure change. The mapping relationship between pressure and sound intensity is a linear function S=20*P+60, where S is the sound intensity (dB) and P is the pressure (N / cm²). The dynamic adjustment of the acoustic levitation force in the range of 0-0.5N is achieved, and the response time is less than 100ms, keeping the levitation point position error within ±0.5mm.

[0022] The pressure distribution data between various parts of the patient's body and the stretcher are collected in real time by arranging a pressure sensor array on the surface of the stretcher, and the pressure data is transmitted to the ultrasonic standing wave control system. The control system controls the phased array ultrasonic transducer array to adjust the phase and amplitude of the ultrasonic wave according to the pressure distribution, so that the focus of the ultrasonic standing wave field is the dynamic target area, and the acoustic levitation force is adjusted according to the pressure value.

[0023] The pressure distribution data between various parts of the patient's body and the stretcher are acquired by a high-precision pressure sensor array on the surface of the stretcher; a convolutional neural network with a U-Net architecture is used to perform real-time analysis on the pressure distribution data to obtain the center coordinates and pressure values ​​of the pressure abnormality area and the most seriously injured part; a pressure distribution heat map is generated according to the center coordinates and pressure values; a genetic algorithm is used to calculate the focal position and sound intensity required for the ultrasonic standing wave field according to the pressure distribution heat map; the sound field distribution is calculated using the sound pressure acoustic equation to obtain the driving parameters of the phased array ultrasonic transducer array; the driving parameters are transmitted to the phased array ultrasonic transducer array to dynamically adjust the phase and amplitude of each transducer; if the pressure value of the injured part is greater than a preset threshold, the transducer output power is adjusted according to the relationship F=k / P, where F is the acoustic levitation force, P is the pressure value, and k is a constant; the acoustic levitation force is changed by adjusting the output power of the transducer to achieve decompression of the injured part.

[0024] Specifically, a high-precision pressure sensor array is arranged on the surface of the stretcher, and the pressure distribution data between various parts of the patient's body and the stretcher are collected in real time through the data acquisition module. The collected analog signals are converted into digital signals, and the pressure data is transmitted to the central processing unit through a high-speed data transmission interface. The convolutional neural network with the U-Net architecture is used to analyze the pressure distribution data in real time, identify the abnormal pressure areas and the most seriously injured parts, calculate the center coordinates and pressure values ​​of these areas, generate the pressure distribution heat map, and pass the analysis results to the ultrasonic standing wave control system. Based on the pressure distribution heat map, the ultrasonic standing wave control system uses a genetic algorithm to calculate the optimal focal position and sound intensity required for the ultrasonic standing wave field, uses the sound pressure acoustic equation to calculate the sound field distribution, and generates the driving parameters of the phased array ultrasonic transducer array, including the phase and amplitude values ​​of each transducer. The control system transmits the optimized driving parameters to the phased array ultrasonic transducer array, dynamically adjusts the phase and amplitude of each transducer, so that the focus of the ultrasonic standing wave field is accurately aligned with the target area, and adjusts the acoustic levitation force in real time according to the pressure value. The inverse function F=k / P is used to describe the relationship between the pressure value and the acoustic levitation force, where F is the acoustic levitation force, P is the pressure value, and k is a constant. The acoustic levitation force is changed by adjusting the transducer output power to achieve accurate decompression of the injured part. 1,000 high-precision pressure sensors are arranged on the surface of the stretcher, with a sensitivity of 0.1N / cm² and a sampling frequency of 100Hz. The analog signal is converted into a digital signal through a 16-bit ADC, and the data transmission rate reaches 1Gbps. The U-Net network contains 4 layers of downsampling and 4 layers of upsampling, uses the ReLU activation function, inputs a 100x100 pressure matrix, and outputs a heat map of the same size. The network identified three pressure anomaly areas with center coordinates of (25,30), (60,45), and (80,70), corresponding to pressure values ​​of 2.5N / cm², 3.2N / cm², and 1.8N / cm². The genetic algorithm sets the population size to 100, iterates 500 times, and calculates the optimal focal position (62,48) and the required sound intensity of 85dB. In the sound pressure acoustic equation, the air density is set to 1.225kg / m³, the sound speed is 343m / s, and the frequency is 40kHz. The sound field distribution is calculated, and the maximum sound pressure value is 1000Pa. The phased array contains 64 transducers, with a phase adjustment accuracy of 0.1° and an amplitude adjustment range of 0-5W. The k value in the inverse function is set to 10N·cm². When the pressure is 2N / cm², the acoustic levitation force is calculated to be 5N. By adjusting the transducer output power from 1W to 5W, dynamic adjustment of the acoustic levitation force in the range of 1-5N is achieved, with a response time of less than 50ms, the focus position error is kept within ±0.5mm, and the acoustic levitation force control accuracy is 0.1N.

[0025] S105. Monitor the vibration of the stretcher in real time, obtain the vibration frequency and amplitude data through the acceleration sensor, compare them with the preset vibration frequency and amplitude data safety thresholds, and trigger an alarm if the threshold is exceeded. At the same time, input the vibration data into the electromagnetic suspension and ultrasonic suspension control system, and actively suppress the vibration by adjusting the electromagnetic coil current and the ultrasonic standing wave sound intensity.

[0026] The obtained vibration data is analyzed in the frequency domain by using a fast Fourier transform algorithm to obtain the main frequency components and corresponding amplitudes of the vibration; based on the main frequency components and amplitudes, it is determined whether a preset safety threshold is exceeded; if the main frequency components and amplitudes exceed the preset safety threshold, an alarm signal is triggered; the vibration data is processed by using a minimum mean square error adaptive filtering algorithm to extract the main vibration components that need to be suppressed; based on the main vibration components that need to be suppressed, the amplitude and phase of the suppression signal are calculated; the suppression signal is generated and input into an electromagnetic suspension control system and an ultrasonic suspension control system; the electromagnetic suspension control system calculates the compensation current value required by the electromagnetic coil according to the suppression signal; the ultrasonic suspension control system adjusts the driving parameters of the ultrasonic transducer array according to the suppression signal; and active suppression of the stretcher vibration is achieved through the coordinated control of the electromagnetic suspension control system and the ultrasonic suspension control system.

[0027] Specifically, a high-precision three-axis acceleration sensor array is installed at the key position of the stretcher, and the vibration data of the stretcher is obtained in real time through the data acquisition module. The collected analog signal is converted into a digital signal, and the vibration data is transmitted to the central processing unit through a high-speed data transmission interface. The fast Fourier transform algorithm is used to perform frequency domain analysis on the vibration data to extract the main frequency components and corresponding amplitudes of the vibration. The analysis results are compared with the pre-set safety thresholds, with a frequency threshold of 10Hz and an amplitude threshold of 0.5g. If the threshold is exceeded, an alarm signal is triggered, and the over-threshold data is recorded and stored. The minimum mean square error (LMS) adaptive filtering algorithm is used to process the vibration data, extract the main vibration components that need to be suppressed, calculate the amplitude and phase of the suppression signal according to the vibration characteristics, generate the corresponding suppression signal, and input the suppression signal into the electromagnetic suspension and ultrasonic suspension control systems respectively. The electromagnetic suspension control system calculates the compensation current value required by the electromagnetic coil according to the input suppression signal, and adjusts the current intensity and direction of the electromagnetic coil through the PWM driver, which has an operating frequency of 20kHz and a duty cycle accuracy of 0.1%. The ultrasonic suspension control system adjusts the driving parameters of the ultrasonic transducer array according to the suppression signal to change the sound intensity distribution of the ultrasonic standing wave field. The two systems achieve active suppression of stretcher vibration through a collaborative control strategy. The system response time is less than 10ms and the vibration suppression rate reaches 90%. 16 triaxial acceleration sensors are installed on the stretcher with a sensitivity of 100mV / g and a sampling frequency of 1kHz. The digital signal is converted through a 16-bit ADC with a data transmission rate of 10Mbps. The fast Fourier transform uses a 1024-point FFT with a frequency resolution of 0.98Hz. The main vibration frequencies are identified as 5.2Hz and 12.8Hz, with amplitudes of 0.3g and 0.6g respectively. After comparing with the preset threshold, the 12.8Hz vibration triggers an alarm. The LMS algorithm uses a 32nd-order FIR filter, with a step size parameter of μ=0.01 and a convergence time of 20ms. The calculated suppression signal amplitude is 0.58g and the phase delay is 45°. The electromagnetic suspension system uses 8 electromagnetic coils with a maximum current of 5A. The PWM driver operates at 20kHz, with a duty cycle adjustment range of 10%-90% and a resolution of 0.1%. The ultrasonic suspension system contains 64 transducers with a frequency of 40kHz and a maximum sound pressure of 1000Pa. The collaborative control strategy uses a fuzzy PID algorithm. The fuzzy rule base contains 49 rules, and the PID parameters are Kp=0.8, Ki=0.2, and Kd=0.05. The overall response time of the system is 8.5ms, and the vibration suppression rate in the 5-20Hz frequency band is 92%. The vibration amplitude at 12.8Hz is reduced from 0.6g to 0.048g.

[0028] S106. Comprehensively analyze the changing trends of the patient's physiological data, changes in pressure distribution, and stretcher vibration to predict the progression of the patient's condition. When the prediction results indicate that the condition may worsen, adjust the electromagnetic suspension and ultrasonic suspension strategies in advance to increase protection of areas susceptible to secondary injuries, and issue early warnings to medical staff to provide reference for subsequent treatment decisions.

[0029] Acquire the patient's physiological data, pressure distribution data and stretcher vibration data; according to the patient's physiological data, the pressure distribution data and the stretcher vibration data, use the Kalman filter algorithm to perform time synchronization and spatial alignment processing on the data to obtain a unified multi-dimensional data stream; use the long short-term memory network to perform time series analysis on the multi-dimensional data stream to obtain time dynamic characteristics, and use the convolutional neural network to extract spatial characteristics of the pressure distribution data to obtain spatial characteristics; merge the time dynamic characteristics with the spatial characteristics to construct a comprehensive representation of the patient's state; use a multi-layer perceptron to analyze the comprehensive representation to obtain the patient's condition deterioration probability value and the expected time of occurrence; if the condition deterioration probability value exceeds a preset threshold, trigger a warning signal; calculate new electromagnetic coil current distribution parameters and ultrasonic standing wave field parameters according to the warning signal and the condition deterioration probability value; adjust the electromagnetic suspension of the stretcher through the new electromagnetic coil current distribution parameters and the ultrasonic standing wave field parameters.

[0030] Specifically, the patient's physiological data, pressure distribution and stretcher vibration information are collected through multi-source sensors, and the Kalman filter algorithm is used to synchronize and align different types of data in time and space to form a unified multi-dimensional data stream, and the data is preprocessed, including denoising, normalization and feature extraction. The long short-term memory network (LSTM) is used to perform time series analysis on the preprocessed multi-dimensional data stream to extract the temporal dynamic characteristics of the patient's state. At the same time, the convolutional neural network (CNN) is used to extract spatial features of the pressure distribution data. The two features are fused through the attention mechanism to construct a comprehensive representation of the patient's state. Based on the fused features, the multi-layer perceptron (MLP) model with dropout regularization technology is used to predict the patient's disease development trend, output the probability value of disease deterioration and the expected time of occurrence, and trigger an early warning signal when the prediction result exceeds the preset threshold. According to the early warning signal and the predicted disease development trend, the pre-trained deep reinforcement learning model is used to automatically adjust the electromagnetic suspension and ultrasonic suspension strategies, and calculate the new electromagnetic coil current distribution and ultrasonic standing wave field parameters. At the same time, the U-Net algorithm is used to analyze the pressure distribution map, identify high-risk areas susceptible to secondary injuries, increase the support strength of these areas in a targeted manner, and send warning information and disease prediction reports to medical staff through the communication module. Multi-source sensors include electrocardiogram (sampling rate 1000Hz), blood oxygen saturation (sampling rate 1Hz), pressure sensor array (64x64, sampling rate 100Hz) and accelerometer (sampling rate 200Hz). The Kalman filter algorithm uses a state space model, observation noise covariance R=0.1, process noise covariance Q=0.01, to achieve data synchronization and alignment. The LSTM network contains 2 layers, each with 128 units. The CNN uses 4 layers of convolution, with a convolution kernel size of 3x3, and the number of channels is 32, 64, 128, and 256 respectively. The 2 layers are fully connected, with 1024 and 512 nodes respectively. The attention mechanism uses scaled dot-product attention, which integrates LSTM and CNN features. The MLP contains 3 hidden layers, with 512, 256, and 128 nodes, respectively, and a dropout rate of 0.5. The threshold for warning of worsening of the condition is set to 0.7. The deep reinforcement learning model adopts a dual DQN structure, with a state space dimension of 256, and an action space of discretized current and sound intensity adjustment values, each with 16 levels. The U-Net contains 4 downsampling and 4 upsampling, uses the ReLU activation function, and the last layer uses the Sigmoid function to output a probability map. The model was pre-trained on a historical data set (10,000 cases) with an accuracy of 95%. It is executed every 100ms, with an average processing time of 87ms, a prediction lead time of 30 minutes, a suspension force adjustment accuracy of 0.1N, and a risk area identification accuracy of 92%.

[0031] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations. In addition, the various different embodiments of the present invention can also be combined arbitrarily, as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A magnetic suspension shock-absorbing road condition perception and active shock-absorbing method for an ambulance, characterized in that: The method comprises: Identify the patient's injured part and injury degree information, establish a mapping table between injury and current intensity in advance, and adjust the current intensity of the electromagnetic coils in different areas under the stretcher according to the electromagnetic coil current intensity control parameters corresponding to the mapping table between injury and current intensity, so as to change the local magnetic field distribution and affect the repulsive force between the magnetic rings in the neodymium iron suspension device; The fuzzy control algorithm is used to dynamically calculate the current intensity control amount of each electromagnetic coil according to the patient's injured part, injury degree information and the condition of other uninjured parts of the body, so as to apply a small support force to the seriously injured part and maintain normal support force for the uninjured part. Identify the stretcher position of the injured part, and generate a directional focused ultrasonic standing wave field through the phased array ultrasonic transducer array just below the sensitive injured part. Use the principle of acoustic levitation to form a micro-suspension point on the patient's body surface, further reducing local stress and avoiding secondary injury to the fracture site. The pressure sensor array detects the contact pressure distribution between the patient's body parts and the stretcher in real time, feeds the pressure data to the ultrasonic standing wave control system, and dynamically adjusts the focus position and sound intensity of the ultrasonic standing wave field, so that the levitation point is always aimed at the most seriously injured area, and changes the size of the acoustic levitation force as needed; The vibration of the stretcher is monitored in real time. The vibration frequency and amplitude data are obtained through the acceleration sensor. The data is compared with the preset safety threshold of the vibration frequency and amplitude data. If the threshold is exceeded, an alarm is triggered. At the same time, the vibration data is input into the electromagnetic suspension and ultrasonic suspension control system. The vibration is actively suppressed by adjusting the electromagnetic coil current and the ultrasonic standing wave sound intensity. Comprehensively analyze the changing trends of the patient's physiological data, changes in pressure distribution, and stretcher vibration to predict the progression of the patient's condition. When the prediction results show that the condition may worsen, adjust the electromagnetic suspension and ultrasonic suspension strategies in advance, increase the protection of parts that are vulnerable to secondary injuries, and issue early warnings to medical staff to provide a reference for subsequent treatment decisions.

2. The method according to claim 1, characterized in that The patient's injured part and injury degree information are identified, an injury condition and current intensity mapping table is pre-established, and the current intensity of the electromagnetic coils in different areas under the stretcher is adjusted according to the electromagnetic coil current intensity control parameters corresponding to the injury condition and current intensity mapping table, so as to change the local magnetic field distribution to affect the repulsive force between the magnetic rings in the neodymium iron suspension device, including: Obtain patient images, use the YOLO algorithm to locate the patient's injured parts, and classify and evaluate the severity of the injury based on the positioning results; Determine the current intensity control parameters required for the electromagnetic coil according to the injury severity assessment result and a pre-established injury and current intensity lookup table mapping; According to the location information of the injured part of the patient and the size of the stretcher, the coordinates of the specific area under the stretcher corresponding to the injured part are calculated, and the current intensity is distributed to the electromagnetic coils around the area; Finite element analysis was performed on the current intensity distribution results to solve Maxwell's equations and obtain the magnetic flux density in different areas under the stretcher. The magnetic flux density distribution is used to perform multi-body dynamics simulation of the suspension system. The equilibrium position between the magnetic rings is determined through iterative calculation, and the suspension height distribution of each point on the stretcher is obtained.

3. The method according to claim 1, characterized in that The fuzzy control algorithm is used to dynamically calculate the current intensity control amount of each electromagnetic coil according to the patient's injured part, injury degree information and the condition of other uninjured parts of the body, so as to apply a small support force to the seriously injured part and maintain a normal support force for the uninjured part, including: Obtaining a full-body image of the patient, scanning the full-body image of the patient to obtain the position coordinates and area size of the injured area, and conducting a quantitative assessment of the injured area to determine the numerical value of the degree of injury; According to the injury degree value and the size of the injured area, a Mamdani fuzzy inference system is established, input variables and output variables are set, and the input variables are fuzzified to obtain a fuzzy set of the output variables; According to the fuzzy set of the output variables, a relationship model between the current intensity and the magnetic field intensity is established, the magnetic field intensity is calculated and the magnetic force is obtained, and the size of the supporting force is determined; The PID controller is used to adjust the current of the electromagnetic coils in different areas under the stretcher in real time, changing the local magnetic field distribution, affecting the repulsive force between the magnetic rings in the neodymium iron suspension device, and exerting support force on the injured area; It also includes: collecting the injured part, the degree of injury and the uninjured part of the patient as input variables, and fuzzifying the input variables through a membership function to obtain a fuzzy control quantity; A fuzzy rule base is established, and the fuzzy reasoning mechanism is used to deduce the current intensity control quantity of each electromagnetic coil, and the fuzzy control quantity is converted into an accurate current value. The output current intensity is inversely proportional to the severity of the patient's injury.

4. The method according to claim 3, characterized in that The patient's injured part, injury degree and uninjured part are collected as input variables, and the input variables are fuzzified by a membership function to obtain a fuzzy control quantity; a fuzzy rule base is established, and the current intensity control quantity of each electromagnetic coil is derived by a fuzzy reasoning mechanism, and the fuzzy control quantity is converted into an accurate current value, and the output current intensity is inversely proportional to the severity of the patient's injury, including: Determine the location coordinates and area size of the injured part based on the patient's full body scan image; Obtaining an AIS scale assessment value of the injured part, and determining a severity value of the injured part according to the AIS scale assessment value; Establishing a Mamdani fuzzy inference system, wherein the Mamdani fuzzy inference system includes an injury degree input variable, an injury area input variable, and an electromagnetic coil current intensity output variable; The injury degree input variable and the injury area input variable are converted into fuzzy sets using the membership function of the Mamdani fuzzy inference system; Perform fuzzy reasoning on the fuzzy set by the minimum-maximum synthesis method to obtain the fuzzy set of the output variable of the electromagnetic coil current intensity; Defuzzification of the fuzzy set of the output variable of the electromagnetic coil current intensity is performed using the centroid method; The precise current intensity control amount of each electromagnetic coil is determined by the defuzzification process.

5. The method according to claim 1, characterized in that The stretcher position for identifying the injured part is directly below the sensitive injured part, and a phased array ultrasonic transducer array is used to generate a directional focused ultrasonic standing wave field, and a micro suspension point is formed on the surface of the patient's body using the principle of acoustic levitation, so as to further reduce the local force and avoid secondary injury to the fractured part, including: The YOLOv5 algorithm is used to analyze patient images and identify the location and range of the injured area; According to the position and range information of the injured part, it is mapped to the stretcher coordinate system to obtain the position data of the injured part on the stretcher; According to the position data of the injured part on the stretcher, a genetic algorithm is used to optimize the layout of the phased array ultrasonic transducer array and determine the control parameters of the transducer array; The sound field is numerically simulated based on the control parameters of the transducer array to obtain the sound pressure distribution; According to the sound pressure distribution, the phased array ultrasonic transducer array is driven by a sound field controller to generate a directionally focused ultrasonic standing wave field, thereby forming a micro-suspended point array on the surface of the patient's body.

6. The method according to claim 1, characterized in that The pressure sensor array is used to detect the contact pressure distribution between various parts of the patient's body and the stretcher in real time, and the pressure data is fed back to the ultrasonic standing wave control system to dynamically adjust the focus position and sound intensity of the ultrasonic standing wave field so that the suspension point is always aligned with the most seriously injured area, and the size of the acoustic suspension force is changed as needed, including: Obtain contact pressure data between various parts of the patient's body and the stretcher; Analyze the pressure distribution according to the contact pressure data to obtain the center coordinates and pressure values ​​of the abnormal pressure area and the most seriously injured part; After obtaining the center coordinates and pressure values ​​of the pressure abnormality area and the most seriously injured part, a pressure distribution heat map is generated according to the center coordinates and pressure values; The optimal focal position and required sound intensity of the ultrasonic standing wave field are calculated by using the pressure distribution thermodynamic map, and the optimal focal position and required sound intensity are used to optimize the driving parameters of the ultrasonic transducer array; The driving parameters are transmitted to the ultrasonic standing wave control system, and real-time feedback control is achieved through a PID controller to dynamically adjust the phase and amplitude of the transducer array; According to the real-time data of the pressure sensor and the pre-established mapping relationship between the pressure value and the sound intensity, the sound intensity is adjusted to achieve a dynamic change in the size of the acoustic levitation force; The method also includes: collecting pressure distribution data between various parts of the patient's body and the stretcher in real time through a pressure sensor array arranged on the surface of the stretcher, transmitting the pressure data to an ultrasonic standing wave control system, and the control system controls a phased array ultrasonic transducer array to adjust the phase and amplitude of the ultrasonic wave according to the pressure distribution, so that the focus of the ultrasonic standing wave field is a dynamic target area, and the acoustic levitation force is adjusted according to the size of the pressure value.

7. The method according to claim 6, characterized in that The pressure sensor array arranged on the surface of the stretcher collects the pressure distribution data between various parts of the patient's body and the stretcher in real time, transmits the pressure data to the ultrasonic standing wave control system, and the control system controls the phased array ultrasonic transducer array to adjust the phase and amplitude of the ultrasonic wave according to the pressure distribution, so that the focus of the ultrasonic standing wave field is a dynamic target area, and the acoustic levitation force is adjusted according to the size of the pressure value, including: The pressure distribution data between various parts of the patient's body and the stretcher are acquired by a high-precision pressure sensor array on the surface of the stretcher; a convolutional neural network with a U-Net architecture is used to perform real-time analysis on the pressure distribution data to obtain the center coordinates and pressure values ​​of the pressure abnormality area and the most seriously injured part; a pressure distribution heat map is generated according to the center coordinates and pressure values; a genetic algorithm is used to calculate the focal position and sound intensity required for the ultrasonic standing wave field according to the pressure distribution heat map; the sound field distribution is calculated using the sound pressure acoustic equation to obtain the driving parameters of the phased array ultrasonic transducer array; the driving parameters are transmitted to the phased array ultrasonic transducer array to dynamically adjust the phase and amplitude of each transducer; if the pressure value of the injured part is greater than a preset threshold, the transducer output power is adjusted according to the relationship F=k / P, where F is the acoustic levitation force, P is the pressure value, and k is a constant; the acoustic levitation force is changed by adjusting the output power of the transducer to achieve decompression of the injured part.

8. The method according to claim 1, characterized in that: The stretcher vibration is monitored in real time, and the vibration frequency and amplitude data are obtained by an acceleration sensor, and compared with the preset vibration frequency and amplitude data safety thresholds. If the threshold is exceeded, an alarm is triggered, and the vibration data is input into the electromagnetic suspension and ultrasonic suspension control system, and the vibration is actively suppressed by adjusting the electromagnetic coil current and the ultrasonic standing wave sound intensity, including: The fast Fourier transform algorithm is used to perform frequency domain analysis on the acquired vibration data to obtain the main frequency components of the vibration and the corresponding amplitudes; According to the main frequency components and amplitudes, determining whether a preset safety threshold is exceeded; If the main frequency component and amplitude exceed the preset safety threshold, an alarm signal is triggered; Processing the vibration data using a minimum mean square error adaptive filtering algorithm to extract the main vibration components that need to be suppressed; Calculating the amplitude and phase of the suppression signal according to the main vibration component that needs to be suppressed; generating the suppression signal, and inputting the suppression signal into the electromagnetic suspension control system and the ultrasonic suspension control system; The electromagnetic suspension control system calculates the compensation current value required by the electromagnetic coil according to the suppression signal; The ultrasonic suspension control system adjusts the driving parameters of the ultrasonic transducer array according to the suppression signal; Active suppression of the stretcher vibration is achieved through coordinated control of the electromagnetic suspension control system and the ultrasonic suspension control system.

9. The method according to claim 1, characterized in that: The comprehensive analysis of the patient's physiological data change trend, pressure distribution change and stretcher vibration is used to predict the progression of the patient's condition. When the prediction results indicate that the condition may worsen, the electromagnetic suspension and ultrasonic suspension strategies are adjusted in advance to increase the protection of the parts that are susceptible to secondary injuries, and an early warning is issued to medical staff to provide reference for subsequent treatment decisions, including: Obtain patient physiological data, pressure distribution data and stretcher vibration data; According to the patient's physiological data, the pressure distribution data and the stretcher vibration data, a Kalman filter algorithm is used to perform time synchronization and spatial alignment processing on the data to obtain a unified multi-dimensional data stream; Performing time series analysis on the multidimensional data stream using a long short-term memory network to obtain time dynamic features, and extracting spatial features from the pressure distribution data using a convolutional neural network to obtain spatial features; fusing the temporal dynamic features with the spatial features to construct a comprehensive representation of the patient's state; Analyzing the comprehensive representation using a multi-layer perceptron to obtain a probability value of the patient's condition worsening and an estimated time of occurrence; If the probability value of the disease worsening exceeds a preset threshold, a warning signal is triggered; Calculating new electromagnetic coil current distribution parameters and ultrasonic standing wave field parameters according to the warning signal and the probability value of disease progression; The electromagnetic suspension of the stretcher is adjusted by the new electromagnetic coil current distribution parameters and the ultrasonic standing wave field parameters.

Images