A method and system for controlling a mobile assisted standing bed for paraplegic patients

Abstract

This application discloses a control method and system for a mobile assisted standing bed for paraplegic patients. The method includes: generating a synchronous lifting control signal; controlling the synchronous lifting of electric telescopic rods in each area according to the synchronous lifting control signal; calculating the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area; generating head deflection and extension coordinated control signals and foot deflection and extension coordinated control signals according to the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head area; controlling the operation of the electric telescopic rod in the head area according to the head deflection and extension coordinated control signal; and controlling the operation of the foot deflection and extension coordinated control signal according to the foot deflection and extension coordinated control signal. This application provides comprehensive and reliable data support for precise control by integrating all types of core input data, including patient physiological parameters, bed mechanical parameters, and environmental interference parameters.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a control method and system for a mobile assisted standing bed for paraplegic patients. Background Technology

[0002] Assisted standing beds have become a key device for improving complications of prolonged bed rest in paraplegic patients (such as muscle atrophy, pressure sores, and deep vein thrombosis). Their core function is to facilitate a smooth transition from a supine to an upright position through the coordinated drive of an electrically operated telescopic rod, providing support for subsequent leg training. Currently, the control technology of existing assisted standing beds has made some progress, enabling basic positional adjustments such as lifting, lowering, and rotation. However, considering the specific physiological needs of paraplegic patients and the mechanical characteristics of the equipment, many technical shortcomings still need to be addressed.

[0003] Existing standing beds mostly use a uniform control algorithm to drive the electric telescopic rods in the head and foot areas, without fully considering the physiological differences between the human head (vulnerable and posture-sensitive) and the feet (large load fluctuations and multi-structure coordination). The head control lacks a deterministic constraint mechanism for cervical spine safety, which can easily lead to excessive stress on the cervical spine due to insufficient control precision; the foot control relies heavily on traditional mechanical formulas for modeling, making it difficult to cope with nonlinear load changes caused by the movement of the movable bed board and the linkage of the connecting plate components, resulting in poor support stability.

[0004] During the operation of the standing bed, the mechanical wear and frictional damping of the electric telescopic rod, environmental temperature and humidity, vibration interference, and differences in the patient's weight, limb posture, and other human body conditions create multi-dimensional coupled interference. Existing control methods often consider a single parameter in isolation (such as focusing only on angle deviation) and have not established a correction mechanism for the multi-field coupling of mechanics, environment, and human body. This results in large angle control errors and telescopic deviations exceeding 1mm, failing to meet the stringent requirements for support precision for paraplegic patients.

[0005] Paraplegic patients rely on standing beds for support, and pressure sores and deep vein thrombosis are major health risks. However, current control technologies do not incorporate health thresholds (such as pressure sore thresholds and blood flow velocity thresholds for thrombosis risk) into the control logic. The control process only focuses on achieving the target motion parameters of the equipment, without establishing a closed-loop adjustment mechanism for health goals. This makes it difficult to dynamically adapt to the differences in health risks among different patients, resulting in limited protective effects.

[0006] In existing control methods, the conversion of motion parameters (angle, extension, acceleration) to motor drive signals (pulse, voltage) often uses fixed formula mapping, without considering the influence of factors such as motor characteristics, electromagnetic interference, and signal transmission delay, resulting in response deviations between the drive signal and the target parameters. At the same time, the coordinated control of the electric telescopic rods in the head and foot areas lacks a unified benchmark, and the coordination errors of angle, extension, and acceleration exceed 0.5°, 1mm, and 0.05m / s², respectively, which can easily cause patients to shift their body position and feel discomfort. Summary of the Invention

[0007] The purpose of this invention is to provide a control method for a movable assisted standing bed for paraplegic patients to at least solve one of the above-mentioned technical problems.

[0008] One aspect of the present invention provides a method for controlling a movable assisted standing bed for paraplegic patients, the method comprising:

[0009] Step 1: Obtain basic patient information and basic bed parameters;

[0010] Step 2: Calculate the overall safe lifting height of the bed based on the patient's basic information and the basic parameters of the bed, and generate a synchronous lifting control signal;

[0011] Step 3: Send the synchronous lifting control signal to the electric telescopic poles in each area to control the synchronous lifting of the electric telescopic poles in each area;

[0012] Step 4: After obtaining the positioning signals from the electric telescopic poles in each area, generate a locking control signal to the electric telescopic pole in the middle area to lock it as the rotation center axis;

[0013] Step 5: Based on the center of gravity balance model, calculate the initial deflection angle, real-time extension and retraction amount and motion acceleration curve of the electric telescopic rod in the patient's head area, and calculate the initial deflection angle, real-time extension and retraction amount and motion acceleration curve of the electric telescopic rod in the patient's foot area.

[0014] Step 6: Generate a head deflection and extension coordinated control signal based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head area;

[0015] Step 7: Generate a coordinated control signal for foot deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area;

[0016] Step 8: Transmit the head deflection and extension coordination control signal to the electric telescopic rod in the head area, so that the electric telescopic rod in the head area works according to the head deflection and extension coordination control signal;

[0017] Step 9: Transmit the foot deflection and extension coordination control signal to the electric telescopic rod in the foot area, so that the electric telescopic rod in the foot area works according to the foot deflection and extension coordination control signal.

[0018] Optionally, the method for controlling a movable assisted standing bed for paraplegic patients further includes:

[0019] During the operation of the electric telescopic rod in the head area and the electric telescopic rod in the foot area according to the coordinated control signal of head deflection and telescopic extension, the posture information fed back by the posture sensor in the head area and the posture information fed back by the posture sensor in the foot area are acquired in real time.

[0020] The deflection speed pulse duty cycle and extension displacement increment are dynamically adjusted based on the attitude information fed back by the attitude sensors in the head area and the foot area.

[0021] Optionally, the step of calculating the overall safe lifting height of the bed based on the patient's basic information and the bed's basic parameters and generating a synchronous lifting control signal includes:

[0022] Based on the patient's basic information and the bed's basic parameters, the three-dimensional offset of the patient's initial lying center of gravity relative to the geometric center of the bed and the positional offset coefficient are calculated using a dynamic calculation model of center of gravity offset.

[0023] Based on the patient's basic information and the bed's basic parameters, the overall restraint safety factor is calculated using the restraint tension safety factor algorithm.

[0024] Based on the patient's basic information and the bed's basic parameters, the equipment's balance reserve coefficient is calculated using a balance reserve torque calculation model.

[0025] The final safe lifting height is generated based on the body position offset coefficient, the overall restraint safety factor, and the equipment balance reserve factor.

[0026] Optionally, the calculation of the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's head region based on the center of gravity balance model includes:

[0027] Based on the patient's basic information and the bed's basic parameters, the reference center of gravity parameters of the electric telescopic rod in the head region are calculated using the head-cervical spine coupling center of gravity correction algorithm.

[0028] Based on the reference center of gravity parameters of the electric telescopic pole in the head area, the relationship between the rotation angle and the telescopic amount of the electric telescopic pole in the head area is calculated by the rotation-telescopic coordinated phase matching algorithm.

[0029] The initial deflection angle of the head, the real-time extension and retraction amount of the head, and the motion acceleration curve are generated based on the relationship between the rotation angle and extension amount of the electric telescopic rod in the head area.

[0030] Optionally, the calculation of the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area includes:

[0031] Obtain a trained three-input neural network model of the foot;

[0032] Foot features are generated based on the patient's basic information and the bed's basic parameters.

[0033] Foot features are input into the foot three-input neural network model to obtain predicted values ​​for rotation angle, extension and contraction, motion acceleration, and bed board collaborative compensation coefficient.

[0034] The predicted values ​​of rotation angle, extension and retraction, motion acceleration, and bed board coordination compensation coefficient are corrected to obtain the corrected predicted values ​​of rotation angle, extension and retraction, and motion acceleration.

[0035] Motion acceleration curves are generated based on the corrected predicted motion acceleration values; among them, the corrected predicted rotation angle is used as the initial deflection angle of the foot, and the corrected predicted extension and contraction is used as the real-time extension and contraction of the foot.

[0036] Optionally, the foot three-input neural network model includes:

[0037] The input layer contains 18 neurons;

[0038] The number of hidden layers is 3.

[0039] The output layer contains four neurons; wherein,

[0040] The 18 neurons are used to receive the following features:

[0041] Lower limb mass, joint angles between connecting plate 1 and connecting plate 2, joint angles between connecting plate 2 and connecting plate 3, joint angles between connecting plate 3 and connecting plate 4, joint torques between connecting plate 1 and connecting plate 2, joint torques between connecting plate 2 and connecting plate 3, joint torques between connecting plate 3 and connecting plate 4, moving state of movable bed board, moving speed of movable bed board, contact tension of thigh side fixing strap 1 and thigh fixing strap 2, contact tension of calf side fixing strap 1 and calf side fixing strap 2, fixing tension of elastic band on footrest, leg pressure sensor array data, leg blood flow velocity, distance from the long side foot area of ​​the rotating frame to the rotation center axis, maximum load of electric telescopic rod, dynamic deformation coefficient of footrest elastic band, joint damping coefficient of connecting plate assembly;

[0042] The hidden layer consists of a first layer with 64 neurons, a second layer with 32 neurons, and a third layer with 16 neurons.

[0043] Optionally, generating a coordinated control signal for head deflection and extension based on the initial head deflection angle, real-time head extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's head region includes:

[0044] Obtain the dynamic deformation of the headrest and the cervical spine safety correction coefficient at the current moment;

[0045] Based on the dynamic deformation of the head pad at the current moment and the cervical spine safety correction coefficient, the initial head deflection angle, the real-time head extension and contraction amount, and the head motion acceleration are corrected respectively, so as to obtain the corrected head target deflection angle, the corrected head motion acceleration, and the corrected head target extension and contraction amount.

[0046] A coordinated control signal for head deflection and extension is generated based on the corrected head target deflection angle, the corrected head motion acceleration, and the corrected head target extension and contraction.

[0047] Optionally, generating a coordinated control signal for foot deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area includes:

[0048] Obtain a trained, uncorrected neural network for the foot;

[0049] Based on the patient's basic information, the bed's basic parameters, and the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the foot area, a fusion feature is generated.

[0050] The fused features are input into a trained foot-uncorrected neural network to obtain a coordinated control signal for foot deflection and extension.

[0051] Optionally, the foot-uncorrected neural network includes:

[0052] The input layer includes 19 signal-specific feature neurons and 3 motion parameter reference value input terminals;

[0053] A feature adaptation layer is used to generate a 32-dimensional fused feature vector based on the 22-dimensional standardized feature vector of the input layer.

[0054] A signal conversion layer is used to generate a 24-dimensional signal adaptation feature vector based on the 32-dimensional fused feature vector output by the feature adaptation layer.

[0055] The output layer is used to generate servo drive pulse width, linear motor drive voltage, and speed closed-loop control voltage based on the 24-dimensional signal adaptation feature vector.

[0056] This application also provides a control system for a mobile assisted standing bed for paraplegic patients, the control system comprising:

[0057] A parameter acquisition module, which is used to acquire basic patient information and basic bed parameters.

[0058] The lifting height generation module is used to calculate the overall safe lifting height of the bed based on the patient's basic information and the basic parameters of the bed, and generate a synchronous lifting control signal.

[0059] A synchronous lifting control signal sending module is used to send synchronous lifting control signals to the electric telescopic poles in each area to control the synchronous lifting of the electric telescopic poles in each area.

[0060] The locking module is used to obtain the positioning signal from the electric telescopic rods in each area, generate a locking control signal to the electric telescopic rod in the middle area, and lock it as the rotation center axis.

[0061] An initial control quantity generation module is used to calculate the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's head area and the electric telescopic rod in the patient's foot area based on a center of gravity balance model.

[0062] A head region control signal generation module is used to generate a head deflection and extension coordinated control signal based on the initial deflection angle, real-time extension amount and motion acceleration curve of the electric telescopic rod in the patient's head region.

[0063] A foot area control signal generation module is used to generate a foot deflection and extension coordinated control signal based on the initial deflection angle, real-time extension amount and motion acceleration curve of the electric telescopic rod in the patient's foot area.

[0064] A control signal transmitting module is used to transmit a head deflection and extension coordination control signal to the electric telescopic rod in the head area, so that the electric telescopic rod in the head area works according to the head deflection and extension coordination control signal, and to transmit a foot deflection and extension coordination control signal to the electric telescopic rod in the foot area, so that the electric telescopic rod in the foot area works according to the foot deflection and extension coordination control signal.

[0065] The mobile assisted standing bed control method for paraplegic patients proposed in this application solves the problems of single input dimension and vague control target in existing assisted standing bed control algorithms. By integrating all types of core input data, such as patient physiological parameters, bed mechanical parameters, and environmental interference parameters, it provides comprehensive and reliable data support for precise control. Attached Figure Description

[0066] Figure 1 This is a flowchart illustrating a method for controlling a movable assisted standing bed for paraplegic patients according to an embodiment of this application.

[0067] Figure 2 A schematic diagram of the structure of the assistive standing bed for paraplegic patients who can use the mobile assistive standing bed control method of this application;

[0068] Figure 3 A schematic diagram of the support frame structure of the assistive standing bed for paraplegic patients who can use the mobile assistive standing bed control method of this application;

[0069] Figure 4 A schematic diagram of the traction rope structure of the assisted standing bed for paraplegic patients who can use the mobile assisted standing bed control method of this application;

[0070] Figure 5 A schematic diagram of the upright position of the assistive standing bed for paraplegic patients using the mobile assistive standing bed control method of this application.

[0071] Figure label:

[0072] 1. Base frame; 2. Electric telescopic rod; 3. Rotating frame; 4. Handle; 5. Fixed bed board; 6. Movable bed board; 7. Support plate; 13. Support frame; 16. Restraint strap one; 17. Buckle; 18. Restraint strap two; 19. Restraint strap three; 20. Soft pad; 22. L-shaped frame; 23. Connecting plate one; 24. Connecting plate two; 25. Connecting plate three; 26. Connecting plate four; 27. Leg frame; 28. Fixing strap one; 29. ​​Fixing strap two; 30. Foot frame; 31. Rotating rod; 32. Rope wheel one; 33. Pull rope one; 34. Rope wheel two; 35. Pull rope two; 36. Motor two. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0074] like Figure 1 The methods for controlling the use of a mobile assisted standing bed for paraplegic patients shown include:

[0075] Step 1: Obtain basic patient information and basic bed parameters;

[0076] Step 2: Calculate the overall safe lifting height of the bed based on the patient's basic information and the basic parameters of the bed, and generate a synchronous lifting control signal;

[0077] Step 3: Send the synchronous lifting control signal to the electric telescopic poles in each area to control the synchronous lifting of the electric telescopic poles in each area;

[0078] Step 4: After obtaining the positioning signals from the electric telescopic poles in each area, generate a locking control signal to the electric telescopic pole in the middle area to lock it as the rotation center axis;

[0079] Step 5: Based on the center of gravity balance model, calculate the initial deflection angle, real-time extension and retraction amount and motion acceleration curve of the electric telescopic rod in the patient's head area, and calculate the initial deflection angle, real-time extension and retraction amount and motion acceleration curve of the electric telescopic rod in the patient's foot area.

[0080] Step 6: Generate a head deflection and extension coordinated control signal based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head area;

[0081] Step 7: Generate a coordinated control signal for foot deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area;

[0082] Step 8: Transmit the head deflection and extension coordination control signal to the electric telescopic rod in the head area, so that the electric telescopic rod in the head area works according to the head deflection and extension coordination control signal;

[0083] Step 9: Transmit the foot deflection and extension coordination control signal to the electric telescopic rod in the foot area, so that the electric telescopic rod in the foot area works according to the foot deflection and extension coordination control signal.

[0084] In this embodiment, the patient's basic information includes the patient's weight parameter m, target upright angle θ, initial lying pressure distribution data P(i,j) (i=1,2,3; j=1,2,3, unit: N) collected by the 3×3 pressure sensor array at the bottom of the bed, the pretension data F1 (shoulder), F2 (chest), F3 (lumbar and hip) of the restraint components 1 (restraint belt 16, restraint belt 2 18, restraint belt 3 19), and the patient's heart rate variability (HRV).

[0085] Basic bed parameters include the maximum load threshold F of the three sets of electric telescopic rods. max The maximum overturning moment M of the bed rotation center axis (electric telescopic rod in the middle area) max The distance L from the geometric center of the bed to the axis of rotation.

[0086] In this embodiment, the step of calculating the overall safe lifting height of the bed based on the patient's basic information and the bed's basic parameter information and generating a synchronous lifting control signal includes:

[0087] Based on the patient's basic information and the bed's basic parameters, the three-dimensional offset of the patient's initial lying center of gravity relative to the geometric center of the bed and the positional offset coefficient are calculated using a dynamic calculation model of center of gravity offset.

[0088] In this embodiment, the calculation of the three-dimensional offset of the patient's initial lying center of gravity relative to the geometric center of the bed and the positional offset coefficient using the dynamic calculation model of center of gravity offset includes:

[0089] The formula for calculating the pressure distribution entropy value S (reflecting the uniformity of the patient's lying position) is:

[0090] ;

[0091] Wherein, P(i,j) represents the initial lying pressure distribution data of the patient collected by the 3×3 pressure sensor array at the bottom of the bed;

[0092] If S≥0.8, output a body position fine-tuning prompt signal; if S<0.8, proceed to the next calculation step.

[0093] Establish a three-dimensional coordinate system with the geometric center of the bed as the origin (X-axis along the head-to-foot direction, Y-axis along the width of the bed, Z-axis perpendicular to the bed surface), and calculate the three-dimensional coordinates of the patient's center of gravity (X). c ,Y c Z c The formula is:

[0094] ;

[0095] ;

[0096] ;

[0097] Wherein, P(i,j) is the initial lying pressure distribution data of the patient collected by the 3×3 pressure sensor array at the bottom of the bed; X(i,j) and Y(i,j) are the coordinates of the (i,j)th pressure sensor, g is the gravitational acceleration, and h0 is the initial thickness of the bed cushion;

[0098] Calculate the three-dimensional offset of the center of gravity (ΔX, ΔY, ΔZ) = (|X c |, |Y c |, |Z c -h0|), body position offset coefficient K p The formula is:

[0099] ;

[0100] Among them, L X For bed length, L Y For bed width, K Р ∈[0.9,1.1];

[0101] Based on the patient's basic information and the bed's basic parameters, the overall restraint safety factor is calculated using the restraint tension safety factor algorithm.

[0102] In this embodiment, based on the patient's basic information and the bed's basic parameters, the overall restraint safety factor is calculated using a restraint tension safety factor algorithm, including:

[0103] Preset safety tension thresholds for each part (Shoulder, 50N≤) ≤80N), (Chest, 60N≤) ≤90N), (Lower back and hips, 70N≤) ≤100N), calculate the tension safety factor for each part:

[0104] ; ; ;

[0105] The overall restraint safety factor is calculated using a weighted fusion method (shoulder weight 0.3, chest weight 0.4, waist and hip weight 0.3):

[0106] ;in, In this embodiment, if Ks <0.7 indicates insufficient restraint, K s >1.3 Determined to be too tight;

[0107] Based on the patient's basic information and the bed's basic parameters, the equipment's balance reserve coefficient is calculated using a balance reserve torque calculation model.

[0108] In this embodiment, based on the patient's basic information and the bed's basic parameters, the equipment's balance reserve coefficient is calculated using a balance reserve torque calculation model, including:

[0109] Calculate the overturning moment when the lifting height H is reached. (Caused by the patient's weight and shift in center of gravity):

[0110] ;

[0111] Calculate the equilibrium reserve coefficient K b The formula is:

[0112] ;

[0113] Among them, M max The maximum anti-overturning moment is the center axis of bed rotation (electric telescopic rod in the middle area); in this embodiment, K is required to be... b ≥0.35 (ensuring no tipping during rotation), the greater the lifting height H, the greater the M overturn The larger K is b The smaller;

[0114] The final safe lifting height is generated based on the body position offset coefficient, the overall restraint safety factor, and the equipment balance reserve factor.

[0115] In this embodiment, the final safe lifting height is generated based on the body position offset coefficient, the overall restraint safety factor, and the equipment balance reserve factor, including:

[0116] Calculate the foundation lift height H base (Determined by the minimum space required for rotation and the distance from the hospital bed):

[0117] ;

[0118] Where D is the distance between the bed and the patient bed (maximum value is 5 meters), H base ∈[0.15m, 0.45m];

[0119] Calculate the dynamic correction coefficient K d (Combined with the patient's heart rate variability, reflecting the patient's level of anxiety):

[0120] ;

[0121] Solve for the final safe lifting height H:

[0122] ;

[0123] Output the final safe lifting height H and generate a synchronous lifting control signal (PWM control signal).

[0124] In this embodiment, after the pole is raised to its designated position (height sensor detection H error ≤ ±0.01m), a high-level height locking signal is output to the three sets of electromagnetic locks on the electric telescopic pole; simultaneously, the calculated K is... p K s K b Parameters such as H and HRV are stored in the patient's personalized parameter library as baseline data for subsequent use.

[0125] In this approach, the present application deeply couples the patient's dynamic body position (quantified by pressure entropy value and three-dimensional center of gravity offset), restraint tension safety (weighted fusion of tension coefficients of multiple sites), and equipment balance reserve (quantification of tipping risk), and incorporates dynamic correction of the patient's heart rate variability. This mechanism solves the risk of lifting and slipping caused by initial body position deviation and improper restraint tension in completely paralyzed patients, and significantly improves the adaptability and safety of different scenarios and different patients.

[0126] In this embodiment, the patient's basic information may also include the following:

[0127] Head mass percentage k h (k) h =0.08), target upright angle θ;

[0128] Basic bed parameters may also include the following parameters: headrest elastic modulus E h Head restraint strap (elastic headband) pretension F h Cervical spine safety pressure threshold P c The distance L from the head region of the long side of the rotating frame to the rotation center axis h The initial pitch angle φ0 and roll angle ψ0 are acquired by the head posture sensor.

[0129] In this embodiment, based on the patient's basic information and the bed's basic parameters, the reference center of gravity parameters of the electric telescopic rod in the head region are calculated using a head-cervical spine coupling center of gravity correction algorithm, including:

[0130] Calculate the head mass m h =k h ×m, based on the head pad elastic modulus E h The deformation offset Δh of the head pad under head pressure is calculated using the formula of elasticity mechanics. The formula is as follows:

[0131] ;

[0132] Among them, S h A represents the contact area between the head and the head pad. h Let g be the moment of inertia of the head pad section, and g be the acceleration due to gravity.

[0133] Establish a coordinate system with the rotation center axis as the origin (X-axis along the length of the bed, Y-axis perpendicular to the bed surface), and calculate the initial coordinates of the head's center of gravity (X... h0 Y h0 The formula is:

[0134] ;

[0135] ;

[0136] Introducing cervical spine safety pressure constraints to correct the head's center of gravity coordinates (X) hc Y hc The formula is as follows:

[0137] ;

[0138] Among them, S c The preset area for cervical spine stress distribution ranges from 0.012 to 0.018 m², and users can set the value as needed.

[0139] In this embodiment, based on the reference center-of-gravity parameters of the electric telescopic pole in the head region, the relationship between the rotation angle and the telescopic amount of the electric telescopic pole in the head region is calculated using a rotation-telescopic cooperative phase matching algorithm, including:

[0140] Set the total rotation time T, divide the target upright angle θ into N iteration steps Δθ=θ / N (Δθ≤0.5°), and the time corresponding to each step Δt=T / N;

[0141] For the i-th iteration (i=1,2,…,N), calculate the current rotation angle θi=i•Δθ, and correct the centroid coordinates (X) based on the head. hc Y hc The target extension S of the electric telescopic pole in the head region is solved by the torque balance equation. h (i), the formula is:

[0142] ;

[0143] Where, k s This refers to the extension stiffness coefficient of the electric telescopic pole.

[0144] Based on the target scaling amount S h (i) Calculate the target rotation angle α of the electric telescopic pole in the head region.h (i) (the angle with the vertical direction), the formula is:

[0145] ;

[0146] Among them, S h (1) is the scaling factor for the first iteration, i.e. the initial scaling factor.

[0147] In this embodiment, the generation of the initial deflection angle of the head, the real-time extension / retraction amount of the head, and the motion acceleration curve based on the relationship between the rotation angle and extension / retraction amount of the electric telescopic rod in the head region includes:

[0148] Calculate attitude deviation Δφ=|φ(i)-φ0|, Δψ=|ψ(i)-ψ0|;

[0149] in, The current pitch angle of the head posture sensor; The tilt angle of the head posture sensor at the current moment;

[0150] If Δφ>1° or Δψ>1°, then the scaling factor in the (i+1)th iteration is corrected using the following formula. :

[0151] ;

[0152] in, Let be the scaling factor for the i-th iteration;

[0153] The initial rotation angle α of the electric telescopic rod in the head region after iterative convergence is... h (1) Initial expansion amount S h (1) The target rotation angle sequence for each iteration step [α] h (1),α h (2),…,α h (N)], target scaling sequence [S] h (1),S h (2),…,S h The (N) output is sent to the controller as the reference control data for the synchronous rotation and extension of the electric telescopic rod in the head area; at the same time, it outputs head posture and cervical spine pressure monitoring signals to ensure that the head posture deviation is ≤1.5° and the cervical spine pressure is within ≤1.5° during the standing process. <Pc。

[0154] In this embodiment, based on the iteratively obtained stretch sequence [S] h (1),S h (2),…,S h The motion acceleration curve a is obtained by fitting the time series [0, Δt, 2Δt, ..., (N-1)Δt] and the corresponding time series [0, Δt, 2Δt, ..., (N-1)Δt]. h(t).

[0155] Using the above technical solution, this application pioneers a three-level control mechanism: head and cervical spine coupling + rotation and extension phase matching + posture closed-loop compensation. Among them, the head and cervical spine coupling center of gravity correction algorithm, rotation-extension collaborative phase matching algorithm, and head posture closed-loop compensation algorithm take into account head-specific factors such as head mass ratio, cervical spine safety pressure, and head pad deformation, and deeply couple them with the rotation-extension control of the electric telescopic pole. This perfectly adapts to the structural characteristics of the electric telescopic pole in the head area, solves the core problems such as cervical spine pressure and posture instability caused by head center of gravity shift during standing, and significantly improves the safety and accuracy of head support.

[0156] In this embodiment, the calculation of the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area includes:

[0157] Obtain a trained three-input neural network model of the foot;

[0158] Foot features are generated based on the patient's basic information and the bed's basic parameters.

[0159] Foot features are input into the foot three-input neural network model to obtain predicted values ​​for rotation angle, extension and contraction, motion acceleration, and bed board collaborative compensation coefficient.

[0160] The predicted values ​​of rotation angle, extension and retraction, motion acceleration, and bed board coordination compensation coefficient are corrected to obtain the corrected predicted values ​​of rotation angle, extension and retraction, and motion acceleration.

[0161] In this embodiment, the predicted values ​​of rotation angle, extension / twist amount, motion acceleration, and bed board collaborative compensation coefficient are corrected to obtain the corrected predicted values ​​of rotation angle, extension / twist amount, and motion acceleration, including:

[0162] Based on the Smove correction parameters according to the moving state of the movable bed board 6:

[0163] If S_move=0 (the movable bed board 6 is in a supported state, providing support for the lower limbs), the corrected parameters are:

[0164] ;

[0165] ;

[0166] ;

[0167] Where, α f_pred S is the predicted value of the rotation angle;f_pred This is the predicted value for the expansion / contraction amount; a f_pred For predicted acceleration; F thigh F calf F foot分别 The tension of the thigh-side fixing strap 28 and the fixing strap 29, the tension of the calf-side fixing strap 28 and the fixing strap 29, and the tension of the elastic band on the tripod 30; max (P leg (x,y) represents the maximum pressure value collected by the leg pressure sensor array; P ulcer The pressure threshold for leg pressure ulcers (1333 Pa corresponding to 32 mmHg); F thr _ sum The thigh and calf restraint tension threshold is set to 200N; M j_thr The joint torque safety threshold for the connecting plate assembly is set to 50 N•m.

[0168] The final corrected rotation angle α f Expansion / Contraction S f , acceleration a f The signal is converted into a PWM drive signal (PWM duty cycle is linearly related to extension / rotation angle, duty cycle change rate is linearly related to acceleration, duty cycle range 5%-95%), and output to the controller of the electric telescopic rod 2 in the foot area; simultaneously, the motion acceleration curve is output (composed of a at each time step a). f (forming a continuous curve), which, together with the rotation angle sequence and the extension / retraction sequence, is stored as reference control data;

[0169] In this embodiment, the foot three-input neural network model includes:

[0170] The input layer contains 18 neurons;

[0171] The hidden layers consist of three layers (64 neurons in the first layer, 32 neurons in the second layer, and 16 neurons in the third layer); the activation function used is ELU (Exponential Linear Unit, to avoid the vanishing gradient problem).

[0172] The output layer contains four neurons; each outputs:

[0173] Rotation angle prediction value α f_pred (Predicted angle between the electric telescopic rod 2 in the foot area and the vertical direction); Predicted telescopic amount S f_pred (Predicted change in length of the electric telescopic pole 2 in the foot area); Predicted acceleration a f_pred (Predicted acceleration of the telescopic movement of the electric telescopic rod 2 in the foot area); Bed board cooperative compensation coefficient k comp(Used to dynamically adapt the movement state of the movable bed board 6, with a value range of [0.9, 1.2]); where,

[0174] The 18 neurons are used to receive the following features:

[0175] Lower limb mass (ml) (referring to the total mass of the patient's thigh, calf, and foot, converted based on human physiological standards and patient weight), joint angle θ between connecting plate one and connecting plate two. j1 (Core motion angle parameters of the connecting plate assembly), joint angle θ between connecting plate two and connecting plate three j2 (The angle of motion at the connection between the thigh and calf affects the leg's supporting posture), the joint angle θ between the third and fourth connecting plates. j3 (Corresponding to the movement angle of the lower leg and the connection point of the foot, adapting to the fixed posture of the foot), the joint torque M of connecting plate one and connecting plate two. j1 (This joint torque data reflects the load supported by the upper thigh), the joint torque M between connecting plate two and connecting plate three. j2 (This joint torque data reflects the load at the connection between the thigh and lower leg.) The joint torque M between connecting plate three and connecting plate four. j3 (The torque data of this joint reflects the load at the connection between the lower leg and the foot), and the movement state S of the movable bed board. move (Binary value, 0 = movable bed board is in supported state, 1 = movable bed board is in detached / moving state), movable bed board moving speed v move (Refers to the real-time speed at which the movable bed board moves along the L-shaped slide 12, affecting the amplitude of sudden changes in foot load), and the contact tension F between thigh side fixing strap one and thigh fixing strap two. thigh (Parameters of restraint component two), the tension F between the first and second calf side fixation straps. calf (Suitable for calf compression load), the tension F of the elastic band on the tripod. foot (Reflecting the tightness of foot fixation), leg pressure sensor array data P leg (x, y) (x, y are the two-dimensional coordinates of the contact plane between the leg and the leg support 27, corresponding to the real-time pressure value at the coordinate point), leg blood flow velocity v blood (Refers to the real-time velocity of venous blood flow in the patient's lower extremities, a core parameter for deep vein thrombosis prevention), and the distance L from the long side foot area of ​​the rotating frame to the rotation center axis. f (Refers to the straight-line distance from the rotation center of the top of the electric telescopic pole 2 in the middle area to the connection point of the top of the electric telescopic pole 2 in the foot area), the maximum load F of the electric telescopic pole. max (Refers to the maximum tensile / compressive force that a single electric telescopic pole 2 can stably withstand), and the dynamic deformation coefficient k of the tripod elastic band. elastic (Calculated based on elastic band stretching, reflecting the degree of elastic cushioning for foot fixation), joint damping coefficient c of the connecting plate assembly. j(The damping characteristics of all joints from the first connecting plate 23 to the fourth connecting plate 26 affect the smoothness of the electric telescopic rod movement in the foot area.)

[0176] This application integrates a data-driven, health-target-bed-board collaborative neural network control architecture based on motion acceleration prediction. The foot three-input neural network model is the first to use motion acceleration as the core control parameter and optimizes 18 foot-specific parameters, such as lower limb load, bed board movement, and dual health targets, across dimensions through neural networks. This solves the problem of impact or sluggishness caused by the lack of precise control over the acceleration of the electric telescopic rod in the foot area in existing technologies, and significantly improves the accuracy, stability, and health protection effect of foot support.

[0177] In this embodiment, the patient's basic information and the bed's basic parameters also include the following data:

[0178] Head region parameter time series (θ) th Head target deflection angle; S th Head target expansion / contraction; a th (Head motion acceleration);

[0179] Head-specific mechanical device - human body parameters:

[0180] Mechanical parameters of the head area of ​​the electric telescopic pole (wear coefficient λ) h Calculation based on the cumulative running time of the telescopic pole in the head area; friction damping coefficient c h Rated output torque M rated_h );

[0181] Head pad characteristic parameters (elastic modulus Eh; Poisson's ratio μh; initial thickness h) h0 );

[0182] Human head related data (head mass m) h ; Cervical spine safety pressure threshold P c Real-time pitch angle φ(t) and roll angle ψ(t) acquired by head posture sensors; real-time tension F of head restraint strap (elastic headband). h(t) );

[0183] Environmental interference parameters (head region vibration acceleration a) vib_h Ambient temperature T h );

[0184] Real-time feedback data (actual deflection angle θ of the head telescopic rod) act_h Actual expansion / contraction S act_h Real-time load F load_h Temperature T of the telescopic rod winding coil_h );

[0185] In this embodiment, the step of generating a coordinated control signal for head deflection and extension based on the initial head deflection angle, real-time head extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's head region includes:

[0186] Obtain the dynamic deformation of the headrest and the cervical spine safety correction coefficient at the current moment;

[0187] In this embodiment, the dynamic deformation of the head pad at the current moment is obtained by the following formula:

[0188] ;

[0189] m h g is the mass of the head; g is the acceleration due to gravity; μ h Poisson's ratio of the head pad material; E h S represents the elastic modulus of the head pad. h The contact area between the head and the head pad; Let be the dynamic deformation of the head pad at time t;

[0190] In this embodiment, the cervical spine safety correction coefficient is obtained using the following formula:

[0191] ;

[0192] Among them, K c F is the cervical spine safety correction factor (value range [0.85, 1.0]); h (t) represents the real-time tension of the head restraint; P c The safe pressure threshold for the cervical spine; S c The preset area for cervical spine stress distribution ranges from 0.012 to 0.018 m², and users can set the value as needed.

[0193] In this embodiment, the initial head deflection angle, real-time head extension / retraction, and head motion acceleration are corrected based on the dynamic deformation of the head cushion at the current moment and the cervical spine safety correction coefficient, thereby obtaining the corrected target head deflection angle, corrected head motion acceleration, and corrected target head extension / retraction, including:

[0194] The target parameters are corrected through multimodal fusion using the following formula:

[0195] ;

[0196] ;

[0197] ;

[0198] Where, θt h_cal The corrected head target deflection angle; θt hThe original head target deflection angle; T coil_h St is the temperature of the head telescopic rod winding. h_cal St represents the corrected head target extension / retraction amount. h Δh represents the original head target expansion / contraction amount. h (t) represents the dynamic deformation of the head pad at time t; h h0 The initial thickness of the head pad; at h_cal This is the corrected head motion acceleration; at h a is the original head motion acceleration; vib_h The acceleration due to vibration in the head region; a vib_thr The safe threshold for head vibration;

[0199] In this embodiment, generating a coordinated control signal for head deflection and extension based on the corrected head target deflection angle, the corrected head motion acceleration, and the corrected head target extension / retraction includes:

[0200] The deflection angle signal is converted into a servo drive pulse signal P. θh The formula is as follows:

[0201] ;

[0202] Among them, P θh For servo motor drive pulse signal; θt h_final =θt h_cal , which is the final corrected head target deflection angle; the pulse frequency is fixed at 50Hz;

[0203] Based on the extension / retraction signal, a PID control method is used to generate the linear motor displacement control signal V. Sh The formula is:

[0204] ;

[0205] Among them, V Sh K is the displacement control signal for the linear motor. ph K is the proportionality constant (value 3.0); ih K is the integral coefficient (value 0.2); dh ΔS is the differential coefficient (value 0.6); h_final =St h_cal- S act_h， This represents the final expansion / contraction deviation. This is the integral term for the expansion / contraction deviation; This is the differential term of the scaling deviation;

[0206] Acceleration mapping and signal fusion are performed to map the corrected head motion acceleration into a velocity closed-loop control signal V_Ah, as shown in the following formula:

[0207] ;

[0208] Among them, V Ah For speed closed-loop control signal; at h_final =at h_cal S represents the final corrected head motion acceleration. h_final =St h_cal S represents the final corrected head target scaling; act_h This refers to the actual extension / retraction of the head telescopic rod; limiting V. Ah ∈[0.008m / s, 0.08m / s] (to avoid excessive speed causing impact or excessively slow speed affecting response);

[0209] Dynamic weighted fusion of three signals is employed: P during the attitude stabilization phase. θh Weight 0.4, V Sh Weight 0.4, V Ah Weight 0.2; Adjustment phase P θh Weight 0.3, V Sh Weight 0.3, V Ah With a weight of 0.4, a head-coordinated control signal U is generated. h_total ; will coordinate control signal U h_total The output is sent to the electric telescopic pole controller in the head area to drive the telescopic pole to perform the movement.

[0210] This application adopts a head-specific collaborative signal generation mechanism that combines multimodal perception fusion with cervical spine safety constraints. For the first time, it deeply integrates parameters such as head cushion elastic deformation, cervical spine safety pressure, and head posture deviation into the control signal generation. Through dynamic weight fusion, it solves the contradiction between control accuracy and safety caused by the fragility and posture sensitivity of the cervical spine in the head area, and significantly improves the accuracy, safety, and comfort of head support.

[0211] In this embodiment, generating a coordinated control signal for foot deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area includes:

[0212] Obtain a trained, uncorrected neural network for the foot;

[0213] In this embodiment, the foot-uncorrected neural network includes an input layer, a feature adaptation layer, a signal conversion layer, and an output layer, wherein,

[0214] The input layer includes 19 signal-specific feature neurons and 3 motion parameter reference value input terminals; the total number of neurons and input dimensions is 22, including 19 signal-specific feature neurons (directly connected to the sensor acquisition module) + 3 motion parameter reference value input terminals (3 outputs of the foot three-input neural network model).

[0215] In this embodiment, the 19 signal features are specifically as follows:

[0216] Motor drive characteristics (5 items): rated pulse range of electric telescopic mast servo, rated drive voltage of linear motor, transmission ratio of servo output shaft, pitch of linear motor, and switching frequency of drive module;

[0217] Signal adaptation features (6 items): head area control signal synchronization frequency, movable bed board movement completion mark, current rotation angle of the rotating frame, real-time extension and retraction position of the electric telescopic rod, electromagnetic clutch working status, and universal wheel locking status (0=unlocked, 1=locked).

[0218] Feedback verification features (4 items): Servo feedback of actual deflection angle, linear motor feedback of actual extension / retraction, speed closed-loop feedback of actual speed, and motor operating current;

[0219] Environmental adaptability features (4 items): ambient temperature, power module output voltage, gap between the rotating frame and the telescopic rod, and tension stability value of the restraint components (average tension of the thigh and calf restraint straps).

[0220] In this embodiment, the three motion parameter baseline values ​​are the predicted values ​​of rotation angle, extension amount, and motion acceleration output by the foot three-input neural network model.

[0221] In this embodiment, the feature adaptation layer is used to generate a 32-dimensional fused feature vector based on the 22-dimensional normalized feature vector of the input layer; specifically, the feature adaptation layer is used to eliminate the dimensional differences between motion parameters and signal features to achieve multimodal deep binding.

[0222] In this embodiment, the feature adaptation layer performs feature fusion using the following function:

[0223] ;

[0224] in, The baseline weights for the motion parameters (converging to 0.4 after training). The 19 signal features are dynamically weighted (optimized through training iterations). The activation terms are adapted to the features (μk and σk are the mean and standard deviation of the k-th feature), which can adaptively enhance effective features and suppress interfering features; The reference value for the deflection angle of the electric telescopic pole in the foot area is the rotation angle prediction value output by the foot three-input neural network model, which is the target angle reference for the telescopic pole to complete the rotational motion around the common rotation axis of the bottom end; The reference value for the extension and retraction of the electric telescopic rod in the foot area is the predicted value of the extension and retraction output by the three-input neural network model of the foot, which is the target length reference for the foot telescopic rod to complete the axial extension and retraction movement. The reference value for the motion acceleration of the electric telescopic pole in the foot area is the motion acceleration prediction value output by the foot three-input neural network model, which is the target acceleration reference for the deflection and extension of the telescopic pole.

[0225] In this embodiment, the output of the feature adaptation layer is a 32-dimensional fused feature vector. Each neuron corresponds to a set of composite features after binding "motion parameters-signal features", and the dimension is consistent with the number of neurons in this layer.

[0226] In this embodiment, the signal conversion layer is used to generate a 24-dimensional signal adaptation feature vector based on the 32-dimensional fused feature vector output by the feature adaptation layer;

[0227] Specifically, the signal conversion layer performs the following operations:

[0228] Linear projection dimensionality reduction: using a trainable weight matrix With bias term The formula for converting from 32 dimensions to 24 dimensions is as follows:

[0229] ;

[0230] The 32-dimensional fused feature vector output by the feature adaptation layer;

[0231] Mapping Relationship Solidification: Based on a model trained with over 3000 sets of full-condition data, the mapping relationship between servo motor pulse-deflection angle and motor voltage-extension is solidified. and ;

[0232] Activation constraints: Applying ReLU6 activation and constraining the eigenrangement to adapt the driving characteristics, the formula is as follows: ;

[0233] The final output is a 24-dimensional signal adaptation feature vector. This is provided for use by the output layer.

[0234] The output layer is used to generate the servo drive pulse width, linear motor drive voltage, and speed closed-loop control voltage based on the 24-dimensional signal adaptation feature vector. In this embodiment, the output layer uses a linear activation function.

[0235] Based on the patient's basic information, the bed's basic parameters, and the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the foot area, a fusion feature is generated.

[0236] The fused features are input into a trained foot-uncorrected neural network to obtain a coordinated control signal for foot deflection and extension.

[0237] In this embodiment, the head of this application adopts deterministic mechanical control to avoid the black box risk of neural networks and ensure the safety of the cervical spine core; the feet adopt direct neural network drive to cope with the complex nonlinearity of signal conversion and improve drive efficiency and accuracy; the two work together to form a global optimization of safety benchmark + high-efficiency drive, which significantly improves the simplicity, accuracy and execution efficiency of control.

[0238] In this embodiment, the method for controlling the movable assisted standing bed for paraplegic patients further includes:

[0239] During the operation of the electric telescopic rod in the head area and the electric telescopic rod in the foot area according to the coordinated control signal of head deflection and telescopic extension, the posture information fed back by the posture sensor in the head area and the posture information fed back by the posture sensor in the foot area are acquired in real time.

[0240] The deflection speed pulse duty cycle and extension displacement increment are dynamically adjusted based on the attitude information fed back by the attitude sensors in the head area and the foot area.

[0241] See Figure 2 , Figure 3 , Figure 4 as well as Figure 5 , Figures 2 to 5 The prior art structural diagram of the mobile assisted standing bed control method for paraplegic patients applicable to this application is shown (for details, please refer to patent application number 2025111192056), and is only briefly described here:

[0242] This invention provides a mobile assisted standing bed for paraplegic patients, comprising: a base frame 1, a bed body assembly, and a support assembly.

[0243] The base frame 1 is connected to the bed frame assembly via three sets of electric telescopic rods 2. By driving the three sets of electric telescopic rods 2 to work together, the bed frame assembly can be raised and lowered relative to the base frame 1, or the bed frame assembly can be rotated from a horizontal position to an upright position and then back to a horizontal position.

[0244] The bed frame assembly includes a rotating frame 3; a fixed bed board 5 is fixedly connected to the inside of the rotating frame 3, and a movable bed board 6 is movably connected to the inside of the rotating frame 3; the fixed bed board 5 provides trunk support, and the movable bed board 6 provides leg support; the movable bed board 6 can be moved from a flat position alongside the fixed bed board 5 to the back of the fixed bed board 5 inside the rotating frame 3 via a moving assembly, so that the patient's legs are suspended in the air;

[0245] The support components include a headrest, armrests, and support frame 13 that conform to the physiological structure of the human body; the headrest and armrests are set on the fixed bed board 5, and the support frame 13 is detachably connected to the fixed bed board 5 through a disassembly assembly;

[0246] The support frame 13 includes a torso assembly and a leg assembly. The torso assembly includes a back frame and a first restraint assembly, while the leg assembly includes a connecting plate assembly, a second restraint assembly, and a drive assembly.

[0247] like Figure 1 As shown, the base frame 1 is generally rectangular in shape, providing overall support for the bed frame assembly, the support components mounted on the bed frame assembly, and the patient's body. The rotating frame 3 is also generally rectangular in shape; three sets of electric telescopic rods 2 are symmetrically arranged on both sides of the base frame 1 and the rotating frame 3, corresponding to the head, middle, and foot areas, for a total of six electric telescopic rods 2. The central axes of the three electric telescopic rods 2 on the same side are coplanar, and the bottom ends of the two electric telescopic rods 2 in the same group are coaxial, meaning that the rotation angle of the two electric telescopic rods 2 in the same group is always the same, and they move synchronously on parallel vertical planes.

[0248] Specifically, the bottom ends of the two electric telescopic rods 2 located in the middle area are respectively set at the midpoint of the long sides of the base frame 1, and are fixedly connected to the electric controllers at corresponding positions inside the base frame 1. Their top ends are rotatably connected to the outer wall of the midpoint of the long side of the rotating frame 3. The electric controllers at these positions can control the two electric telescopic rods 2 in the middle group to perform synchronous lifting and lowering movements, and keep their lifting and lowering heights consistent. The two electric telescopic rods 2 in this group cannot rotate around their bottom ends, and their top ends serve as the rotation centers of the long sides of the rotating frame 3, providing reference support.

[0249] The bottom ends of the two electric telescopic rods 2 located in the head area are rotatably connected to the electric controller in the head area inside the base frame 1 via a common rotating shaft, and their top ends are rotatably connected to the outer wall of the long side of the head area of ​​the rotating frame 3. The electric controller at this position can control the synchronous extension and retraction of the two electric telescopic rods 2, and can also control the synchronous rotation of the two electric telescopic rods 2 by controlling the rotation angle of the common rotating shaft. The bottom ends of the two electric telescopic rods 2 located in the foot area are rotatably connected to the electric controller in the foot area inside the base frame 1 via a common rotating shaft, and their top ends are rotatably connected to the outer wall of the long side of the foot area of ​​the rotating frame 3. The electric controller at this position can control the synchronous extension and retraction of the two electric telescopic rods 2, and can also control the synchronous rotation of the two electric telescopic rods 2 by controlling the rotation angle of the common rotating shaft.

[0250] When the three sets of electric telescopic rods 2 are in their initial state, all three sets are vertical and at the same height, while the rotating frame 3 and its inner bed are horizontal. Pressing the raise / lower button on the remote control allows the three sets of electric telescopic rods 2 to rise and fall synchronously on the vertical plane; releasing the button locks the three sets of electric telescopic rods 2 at their current height. By pressing or releasing the raise / lower button, the plane of the rotating frame, its inner bed, and the support frame 13 can be aligned with the surface of the target hospital bed.

[0251] After the patient is transferred to the assisted standing bed and secured, press the standing switch on the remote control to make the three sets of electric telescopic rods 2 work together: first, the entire bed surface is raised to a suitable height for rotation; then, the two electric telescopic rods 2 in the middle area remain locked at their current height, serving as the support for the rotation center axis; the two sets of electric telescopic rods 2 on the head side and foot side work together to deflect and extend towards the middle area around their common rotation axis at their respective bottom ends, jointly driving the rotating frame 3 to rotate with the head raised and feet lowered until the rotating frame 3 and the inner side of the bed rotate to the preset upright state, and all three sets of electric telescopic rods 2 remain locked.

[0252] During the standing process, the coordinated extension and rotation of the three sets of electric telescopic rods 2 can be carried out synchronously, and their respective rotation speed and extension speed need to be calibrated and determined. The rotating frame 3 serves as the connecting carrier between the bed board and the telescopic rods, and its two long sides are rotatably connected to the top of the electric telescopic rods 2 to ensure a smooth rotation process.

[0253] The base frame 1 has a built-in power module that supplies power to all motor drives and control components of the device. After charging, the built-in power module enables the device to operate normally outdoors without an indoor power supply environment. The base frame 1 is equipped with casters with brakes, and a handle 4 is fixedly connected to the front end of the rotating frame 3. Locking the casters fixes the auxiliary lifting bed in place; releasing the casters allows the auxiliary lifting bed to be moved freely using the handle 4 and the casters.

[0254] A fixed bed board 5 is fixedly connected to the inner circumference of the rotating frame 3 corresponding to the patient's torso area. The fixed bed board 5 is fixedly connected to the rotating frame 3 via a fixed shaft and can rise, fall, or rotate synchronously with the rotating frame 3 to provide support for the patient's torso. Support plates 7 are symmetrically fixedly connected to the bottom of the long sides of the rotating frame 3. The inner surfaces of the two support plates 7 are respectively on the same vertical plane as the corresponding inner wall of the rotating frame 3, serving as extension surfaces of the inner wall of the rotating frame 3. A movable bed board 6 is movably connected between the inner surfaces of the rotating frame 3 and the two support plates 7, corresponding to the patient's leg area, via a movable component. The movable bed board 6 provides support for the patient's legs. Initial state The movable bed board 6 is temporarily kept in a parallel and flat position with the fixed bed board 5 via the movable axis. When the movable bed board 6 is in the initial state, the bottom center line of the movable bed board 6 is collinear with the bottom center line of the fixed bed board 5, and the corresponding sides are basically in contact. At this time, the movable bed board 6 can provide support for the patient's leg area. When the rotating frame 3 rotates to be perpendicular to the base frame 1, the three sets of electric telescopic rods 2 are locked. After the patient completes the standing action in an upright position, the moving component is activated. The movable bed board 6 can be moved along the "L" shaped trajectory to the back of the fixed bed board 5 through the moving component, so that the patient's legs are in a suspended state.

[0255] Specifically, a guide plate is fixedly connected to the top corner area of ​​the bottom surface of the movable bed board 6. A motor is mounted on the rear end of the guide plate via a fixing frame. The drive end of the motor passes through the guide plate and is fixedly connected to a gear. The rear end of the gear is rotatably connected to the front end of the guide plate. A rack 11 is meshed with the outer circumference of the gear and is fixedly connected to a groove in the inner wall of the support plate 7. The groove forms four sets of "L"-shaped slide rails. The short arm of the "L" shape is perpendicular to the bed board plane and extends away from the bed board surface. The long arm of the "L" shape extends parallel to the bed board plane towards the head of the bed. The four sets of "L"-shaped slide rails provide support and sliding tracks for the four bottom corners of the movable bed board 6. The rack is an "L" shape that matches the groove. The gear meshes with the rack under the drive of the motor and makes an "L"-shaped movement in the groove. The cross section of the guide plate matches the inner diameter of the groove and can be partially engaged in the groove to make an "L"-shaped movement synchronously. The guide plate restricts the gear in the groove, preventing it from moving away from the groove.

[0256] Two sets of guide plates, motor one, gears, and racks, consisting of moving components, can be installed at opposite corners of the bottom of the movable bed plate 6 to provide diagonal meshing drive. For diagonal areas without moving components, corresponding guide shafts or guide plates can be installed. Motors one at different corner positions are calibrated to maintain synchronous operation, ensuring stable movement of the movable bed plate 6. Alternatively, four sets of guide plates, motor one, gears, and racks, consisting of moving components, can be installed at the four corners of the bottom of the movable bed plate 6 to provide four-corner meshing drive.

[0257] A headrest is fixedly installed at the head position of the fixed bed board 5. The headrest is arc-shaped and is used to support the patient's head. An elastic headband is installed on the headrest to provide support for the patient's head and cervical spine when necessary.

[0258] The upper surface of the bed frame, composed of the fixed bed board 5 and the movable bed board 6, is padded, with corresponding human-shaped recesses. The support frame as a whole is a human-shaped support structure, conforming to the physiological structure of the human body. The human-shaped recesses on the fixed bed board 5 and the movable bed board 6 match the bottom and side contours of the support frame. The bottom end of the support frame is connected to the fixed bed board 5 via a detachable assembly. Specifically, bolts are fixed at the four corners of the bottom end of the support frame 13, and through holes are made at the corresponding locations of the human-shaped recesses on the fixed bed board 5. The bolts pass through the fixed bed board 5 and are threadedly connected to nuts 15. During use, the support frame 13 is fixedly installed on the fixed bed board 5 with bolts and nuts 15 to prevent displacement. After use, the nuts can be loosened to remove the support frame 13 from the fixed bed board 5 for easy cleaning, disinfection, or storage.

[0259] When the support frame 13 is installed on the fixed bed board 5, the support frame 13 can be embedded in the human-shaped recess on the upper surface of the fixed bed board 5 and the movable bed board 6, so that the support frame 13 is basically flush with the upper surface of the fixed bed board 5 and the movable bed board 6.

[0260] The upper part of the support frame 13 is equipped with a restraint component. Specifically, two restraint straps 16 for restraining the shoulders are respectively provided on both sides of the back frame of the support frame 13 corresponding to the upper end of the patient's shoulders, and two restraint straps 28 for restraining the chest are provided on both sides of the back frame of the support frame 13 corresponding to the patient's ribs. Buckles 17 are provided at the connection points of the restraint straps 16 and 28. The ends of the two restraint straps 16 are connected to the upper ends of the two restraint straps 28 via the buckles 17, and the middle connecting ends of the two restraint straps 28 can also be connected via the buckles 17. The restraint straps 16 extend downwards from the patient's shoulders to the chest, and the restraint straps 28 pass under the patient's armpits to the chest. The restraint straps 16 and 28 can tightly restrain the patient's upper torso to the back frame of the support frame 13.

[0261] The lower part of the back frame of the support frame 13 is equipped with a binding strap 3 19 for binding the patient's waist and hips. The binding strap 3 19 is shaped like a "T" belt. The upper ends of the vertical "T" belt are equipped with buckles 17 on both sides, which connect to the horizontal straps on both sides. The binding strap 3 19 provides binding support from the patient's waist and hips, tightly binding the patient's waist and hips to the waist and hip frame of the support frame 13.

[0262] The tightness of the first restraint belt 16, the second restraint belt 18, and the third restraint belt 19 can be adjusted through the buckle 17 to adapt to the restraint support needs of patients of different body types.

[0263] Handrails are correspondingly installed on the fixed bed boards 5 on both sides of the support frame 13. Specifically, an inverted "L"-shaped storage slot is provided on the upper part of the fixed bed board 5. A rotating shaft is fixedly installed on the inner side of the bottom apex of the inverted "L"-shaped storage slot. A rotating block is rotatably connected to the rotating shaft, and an L-shaped frame 22 is rotatably connected to the top of the rotating block. The L-shaped frame 22 initially matches the inverted "L"-shaped storage slot and can be completely stored in the inverted "L"-shaped storage slot. When the L-shaped frame 22 is disengaged from the inverted "L"-shaped storage slot, it can rotate and open and close on an axis perpendicular to the central axis of the rotating block, and it can also be extended and retracted. The L-shaped frames 22 on both sides can be staggered above the chest of the support frame 13, pass under the patient's armpit, and form a barrier support on the lower part of the patient's upper arm and in front of the chest.

[0264] Two sets of connecting plate assemblies are symmetrically arranged on both sides of the central axis of the support frame 13 back frame, including connecting plate 1 23 fixedly connected to both sides of the support frame 13 back frame. The tail end of each connecting plate 1 23 is rotatably connected to the head end of a connecting plate 24, the tail end of each connecting plate 24 is rotatably connected to the head end of a connecting plate 3 25, and the tail end of each connecting plate 3 25 is rotatably connected to the head end of a connecting plate 4 26. The connecting plates 1 23, 24, 35, and 46 are continuously rotatably connected and symmetrically and parallelly arranged on both sides of the support frame 13, providing a transmission basis for the patient's leg training.

[0265] Leg supports 27 are fixedly connected to the inner sides of two connecting plates 24 and two connecting plates 3, respectively, for a total of four leg supports 27, which are used to provide fixed support for the patient's left and right thighs and calves; each leg support 27 is equipped with a restraint component 2; foot supports 30 are rotatably connected to the inner sides of the tail ends of two connecting plates 4 26, for a total of two foot supports 30; each foot support 30 is fixedly equipped with an elastic band for foot fixation.

[0266] Specifically, a first fixing strap 28 is provided around the front side of the leg frame 27, and a second fixing strap 29 is provided around the side or rear side of the leg frame 27. The inner side of the first fixing strap 28 is the barbed side of the hook and loop fastener, and the outer side of the second fixing strap 29 is the rough side of the hook and loop fastener. The inner side of the first fixing strap 28 can be attached to the outer side of the second fixing strap 29 to form a restraint ring around the leg with the leg frame 27.

[0267] The drive assembly includes a rotating rod 31, which passes through the bottom of the support frame 13 and is rotatably connected to the support frame 13. Both ends of the rotating rod 31 are connected to a first rope wheel 32 via an electromagnetic clutch. The outer periphery of the two first rope wheels 32 is fixedly connected to the tail ends of two second connecting plates 24 via a first pull rope 33. The outer periphery of the two first rope wheels 32 is coaxially fixedly connected to a second rope wheel 34. The outer periphery of the two second rope wheels 34 is fixedly connected to the tail ends of two third connecting plates 25 via a second pull rope 35.

[0268] The drive assembly also includes a worm gear fixedly connected to the outer wall of the rotating rod 31; a screw is meshed at the bottom end of the worm gear, and the drive end of the second motor 36 is fixedly connected to the rear end of the screw. The second motor 36 is mounted at the bottom of the support frame 13 via a fixing bracket.

[0269] The electromagnetic clutches on both sides of the rotating rod 31 operate in opposite states, that is, when one electromagnetic clutch is engaged, the other electromagnetic clutch is disengaged; the first rope wheel 32 and the second rope wheel 34 on the same side have opposite winding directions, that is, when the first rope wheel 32 is tightened, the second rope wheel 34 on the same side is relaxed, so that the patient's thigh and calf movements on that side are coordinated.

[0270] The working principle of the leg training structure driven by the drive component in this invention is as follows: the second motor 36 drives the screw to rotate, and the screw meshes with the worm gear on the outer wall of the rotating rod 31, causing the rotating rod 31 to rotate. Through the electromagnetic clutches on both sides of the rotating rod 31, the transmission state connecting the single-sided rope wheel 32 and the second rope wheel 34 can be independently controlled to achieve precise single-leg training. Taking the left thigh lift as an example: the right electromagnetic clutch is disengaged, and the right rope wheel 1 32 and rope wheel 2 34 stop transmission; the left electromagnetic clutch is magnetically locked, and the rotating rod 31 drives the rear rope wheel 1 32 to rotate, tightening the pull rope 1 33; the pull rope 1 33 is connected to the connecting plate 24, pulling the connecting plate 24 to rotate counterclockwise around the connection point of the connecting plate 1 23, thereby driving the left thigh to lift through the leg frame 27; at this time, the rope wheel 2 34 rotates synchronously, releasing the pull rope 2 35 to avoid pulling interference on the lower leg, so that the lower leg can follow the knee and thigh in linkage; similarly, the reverse drive motor 2 36 can reset the left thigh; if the reverse drive motor 2 36 continues, the rope wheel 34 tightens the pull rope 2 35, and the connecting plate 3 25 rotates around the connecting plate 2 24, driving the left lower leg to lift, at which time the pull rope 1 33 is released again, so that the thigh can follow the knee and lower leg in linkage; the right leg training is achieved by switching the magnetic state of the electromagnetic clutch 39, and the operation logic is the same.

[0271] The specific usage process of this invention is as follows: When it is necessary to transfer a patient to this device, firstly, the device is pushed to the target bed using the casters 40 at the bottom of the base frame 1. The brake device of the casters is locked to ensure stability. At this time, the three sets of electric telescopic rods 2 are in an initial inactive state, the fixed bed board 5 and the movable bed board 6 are in a horizontal state, and the support frame 13 is fixed to the fixed bed board 5 by the disassembly assembly of bolts 14 and nuts 15. All restraint components are released, and the soft pads 20 and other soft structures on its surface conform to the physiological curvature of the human body, providing initial support for the patient. The lift / lower switch is activated, and the three sets of electric telescopic rods 2 are activated and raised and lowered synchronously, adjusting the device to a plane with the same height as the target bed. The lift / lower switch is released, and the three sets of electric telescopic rods 2 remain locked in their current state. Medical staff lay out a transfer mat between the two beds, push the patient to the fixed bed board 5 and the movable bed board 6, so that the patient's back is against the support frame 13, and adjust the body position to a comfortable lying position.

[0272] After confirming that the patient's body is properly secured, activate the standing switch. The three sets of electric telescopic rods 2 work in concert: the three sets of electric telescopic rods 2 first raise and lower synchronously to lift the entire device to a suitable height. Then, the two middle electric telescopic rods, acting as support rods at the center of rotation, remain locked. The electric telescopic rods on the head side and foot side rotate relative to each other and extend and retract simultaneously, causing the entire device to rise headfirst and lower feetmost until the standing action is completed. After the standing action is completed, the three sets of electric telescopic rods remain locked in their current state.

[0273] At this time, medical staff can release the brakes on the casters and use handle 4 to push the device to move it flexibly, allowing the patient to temporarily leave the ward environment and go outdoors for training or socializing.

[0274] This application also provides a control system for a mobile assisted standing bed for paraplegic patients, the control system comprising:

[0275] A parameter acquisition module, which is used to acquire basic patient information and basic bed parameters.

[0276] The lifting height generation module is used to calculate the overall safe lifting height of the bed based on the patient's basic information and the basic parameters of the bed, and generate a synchronous lifting control signal.

[0277] A synchronous lifting control signal sending module is used to send synchronous lifting control signals to the electric telescopic poles in each area to control the synchronous lifting of the electric telescopic poles in each area.

[0278] The locking module is used to obtain the positioning signal from the electric telescopic rods in each area, generate a locking control signal to the electric telescopic rod in the middle area, and lock it as the rotation center axis.

[0279] An initial control quantity generation module is used to calculate the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's head area and the electric telescopic rod in the patient's foot area based on a center of gravity balance model.

[0280] A head region control signal generation module is used to generate a head deflection and extension coordinated control signal based on the initial deflection angle, real-time extension amount and motion acceleration curve of the electric telescopic rod in the patient's head region.

[0281] A foot area control signal generation module is used to generate a foot deflection and extension coordinated control signal based on the initial deflection angle, real-time extension amount and motion acceleration curve of the electric telescopic rod in the patient's foot area.

[0282] A control signal transmitting module is used to transmit a head deflection and extension coordination control signal to the electric telescopic rod in the head area, so that the electric telescopic rod in the head area works according to the head deflection and extension coordination control signal, and to transmit a foot deflection and extension coordination control signal to the electric telescopic rod in the foot area, so that the electric telescopic rod in the foot area works according to the foot deflection and extension coordination control signal.

[0283] The technical solution of this application has the following advantages:

[0284] Addressing the vulnerability and posture sensitivity of the head and cervical spine, a three-level control system is employed, consisting of head-cervical spine coupled center of gravity correction, rotation-extension coordinated phase matching, and posture closed-loop compensation. This system incorporates parameters such as dynamic deformation of the headrest and cervical spine safety pressure thresholds to prevent excessive stress on the cervical spine and ensure safe head support. For the feet, a three-input neural network model with 18 input neurons integrates multi-dimensional features such as lower limb load, bed board movement, and leg blood flow velocity to accurately predict rotation angle, extension / extension, and motion acceleration, thus responding to nonlinear load changes and improving support stability.

[0285] Multi-dimensional data fusion strengthens the foundation for safety control: Breaking through the limitations of traditional single-parameter control, it comprehensively integrates patient physiological parameters (weight, heart rate variability, blood flow velocity, etc.), bed mechanical parameters (electric telescopic rod load, connecting plate joint damping, etc.), and environmental interference parameters (temperature, humidity, vibration acceleration, etc.). Through the coupling of three models—dynamic calculation of center of gravity offset, restraint tension safety factor, and balance reserve torque—it generates the optimal safe lifting height. At the same time, it incorporates dynamic correction of patient heart rate variability to eliminate the risk of slippage caused by initial positional deviation and improper restraint, adapting to different patients and scenarios.

[0286] Real-time closed-loop correction and dynamic optimization of control accuracy: Throughout the entire operation of the standing bed, posture information is collected in real time by head and foot posture sensors, and the duty cycle of the deflection speed pulse and the increment of the extension displacement are dynamically adjusted to compensate for errors caused by mechanical wear and environmental interference. In the head control, target parameters are also corrected through multimodal fusion, while the feet continuously optimize the predicted values ​​with the help of a neural network model, which greatly reduces the angle control error and extension deviation, far exceeding the accuracy standards of existing technologies, and avoids body position deviation and physical discomfort.

[0287] Intelligent conversion of drive signals, balancing safety and efficiency: The head adopts deterministic mechanical control logic, which dynamically weights and fuses servo pulses, linear motor displacement and speed closed-loop control signals to avoid the black box risk of neural networks and ensure the safety of the cervical spine core; the feet use an uncorrected neural network with feature adaptation layer and signal conversion layer to accurately convert 22-dimensional fused features into servo drive pulse width and linear motor drive voltage, incorporating factors such as motor characteristics and electromagnetic interference to reduce the response deviation between drive signals and target parameters, while realizing coordinated control of the head and feet and reducing coordination error.

[0288] Integrated health protection goes beyond simple position adjustment: For the first time, health thresholds (pressure ulcer pressure threshold, blood flow velocity threshold for deep vein thrombosis risk) are incorporated into the control logic. In foot control, pressure is monitored in real time through a leg pressure sensor array, and the support status is dynamically adjusted in combination with blood flow velocity parameters, forming a closed-loop adjustment mechanism for health goals. The calculation of the lifting height also takes into account the safety factor of the restraint tension, avoiding health risks caused by excessive or insufficient restraint. This achieves smooth position transitions and effectively prevents complications from prolonged bed rest.

[0289] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A method for controlling a movable assisted standing bed for paraplegic patients, characterized in that, The method for controlling the mobile assisted standing bed for paraplegic patients includes: Step 1: Obtain basic patient information and basic bed parameters; Step 2: Calculate the overall safe lifting height of the bed based on the patient's basic information and the bed's basic parameters, and generate a synchronous lifting control signal; the calculation of the overall safe lifting height of the bed based on the patient's basic information and the bed's basic parameters, and the generation of the synchronous lifting control signal, includes: Based on the patient's basic information and the bed's basic parameters, the three-dimensional offset of the patient's initial lying center of gravity relative to the geometric center of the bed and the positional offset coefficient are calculated using a dynamic calculation model of center of gravity offset. Based on the patient's basic information and the bed's basic parameters, the overall restraint safety factor is calculated using the restraint tension safety factor algorithm. Based on the patient's basic information and the bed's basic parameters, the equipment's balance reserve coefficient is calculated using a balance reserve torque calculation model. The final safe lifting height is generated based on the body position offset coefficient, the overall restraint safety factor, and the equipment balance reserve factor. Step 3: Send the synchronous lifting control signal to the electric telescopic poles in each area to control the synchronous lifting of the electric telescopic poles in each area; Step 4: After obtaining the positioning signals from the electric telescopic poles in each area, generate a locking control signal to the electric telescopic pole in the middle area to lock it as the rotation center axis; Step 5: Based on the center of gravity balance model, calculate the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head area; calculate the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area; the calculation of the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head area based on the center of gravity balance model includes: Based on the patient's basic information and the bed's basic parameters, the reference center of gravity parameters of the electric telescopic rod in the head region are calculated using the head-cervical spine coupling center of gravity correction algorithm. Based on the reference center of gravity parameters of the electric telescopic pole in the head area, the relationship between the rotation angle and the telescopic amount of the electric telescopic pole in the head area is calculated by the rotation-telescopic coordinated phase matching algorithm. The initial deflection angle of the head, the real-time extension and retraction amount of the head, and the motion acceleration curve are generated based on the relationship between the rotation angle and extension amount of the electric telescopic rod in the head area. The calculation of the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area includes: Obtain a trained three-input neural network model of the foot; Foot features are generated based on the patient's basic information and the bed's basic parameters. Foot features are input into the foot three-input neural network model to obtain predicted values ​​for rotation angle, extension and contraction, motion acceleration, and bed board collaborative compensation coefficient. The predicted values ​​of rotation angle, extension and retraction, motion acceleration, and bed board coordination compensation coefficient are corrected to obtain the corrected predicted values ​​of rotation angle, extension and retraction, and motion acceleration. Motion acceleration curves are generated based on the corrected predicted values ​​of motion acceleration; among them, the corrected predicted value of rotation angle is used as the initial deflection angle of the foot, and the corrected predicted value of extension and contraction is used as the real-time extension and contraction of the foot. The foot-based three-input neural network model includes: The input layer contains 18 neurons; The number of hidden layers is 3. The output layer contains four neurons; wherein, The 18 neurons are used to receive the following features: Lower limb mass, joint angles between connecting plate 1 and connecting plate 2, joint angles between connecting plate 2 and connecting plate 3, joint angles between connecting plate 3 and connecting plate 4, joint torques between connecting plate 1 and connecting plate 2, joint torques between connecting plate 2 and connecting plate 3, joint torques between connecting plate 3 and connecting plate 4, moving state of movable bed board, moving speed of movable bed board, contact tension of thigh side fixing strap 1 and thigh fixing strap 2, contact tension of calf side fixing strap 1 and calf side fixing strap 2, fixing tension of elastic band on footrest, leg pressure sensor array data, leg blood flow velocity, distance from the long side foot area of ​​the rotating frame to the rotation center axis, maximum load of electric telescopic rod, dynamic deformation coefficient of footrest elastic band, joint damping coefficient of connecting plate assembly; The hidden layer has 64 neurons in the first layer, 32 neurons in the second layer, and 16 neurons in the third layer. Step 6: Generate a head deflection and extension coordinated control signal based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head area; The process of generating a coordinated control signal for head deflection and extension based on the initial head deflection angle, real-time head extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's head region includes: Obtain the dynamic deformation of the headrest and the cervical spine safety correction coefficient at the current moment; Based on the dynamic deformation of the head pad at the current moment and the cervical spine safety correction coefficient, the initial head deflection angle, the real-time head extension and contraction amount, and the head motion acceleration are corrected respectively, so as to obtain the corrected head target deflection angle, the corrected head motion acceleration, and the corrected head target extension and contraction amount. A head deflection and extension coordinated control signal is generated based on the corrected head target deflection angle, the corrected head motion acceleration, and the corrected head target extension and retraction. Step 7: Generate a coordinated control signal for foot deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area; The generation of a coordinated control signal for foot deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area includes: Obtain a trained, uncorrected neural network for the foot; Based on the patient's basic information, the bed's basic parameters, and the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the foot area, a fusion feature is generated. The fused features are input into a trained foot-uncorrected neural network to obtain a coordinated control signal for foot deflection and extension. The foot-uncorrected neural network includes: The input layer includes 19 signal-specific feature neurons and 3 motion parameter reference value input terminals; A feature adaptation layer is used to generate a 32-dimensional fused feature vector based on the 22-dimensional standardized feature vector of the input layer. A signal conversion layer is used to generate a 24-dimensional signal adaptation feature vector based on the 32-dimensional fused feature vector output by the feature adaptation layer. The output layer is used to generate servo drive pulse width, linear motor drive voltage and speed closed-loop control voltage based on 24-dimensional signal adaptation feature vectors. Step 8: Transmit the head deflection and extension coordination control signal to the electric telescopic rod in the head area, so that the electric telescopic rod in the head area works according to the head deflection and extension coordination control signal; Step 9: Transmit the foot deflection and extension coordination control signal to the electric telescopic rod in the foot area, so that the electric telescopic rod in the foot area works according to the foot deflection and extension coordination control signal.

2. The method for controlling a movable assisted standing bed for paraplegic patients as described in claim 1, characterized in that, The method for controlling a mobile assisted standing bed for paraplegic patients further includes: During the operation of the electric telescopic rod in the head area and the electric telescopic rod in the foot area according to the coordinated control signal of head deflection and telescopic extension, the posture information fed back by the posture sensor in the head area and the posture information fed back by the posture sensor in the foot area are acquired in real time. The deflection speed pulse duty cycle and extension displacement increment are dynamically adjusted based on the attitude information fed back by the attitude sensors in the head area and the foot area.

3. A control system for a mobile assisted standing bed for paraplegic patients, characterized in that, The control system for the mobile assisted standing bed for paraplegic patients includes: A parameter acquisition module, which is used to acquire basic patient information and basic bed parameters. A lifting height generation module is used to calculate the overall safe lifting height of the bed based on the patient's basic information and the bed's basic parameters, and to generate a synchronous lifting control signal; the calculation of the overall safe lifting height of the bed based on the patient's basic information and the bed's basic parameters, and the generation of the synchronous lifting control signal, includes: Based on the patient's basic information and the bed's basic parameters, the three-dimensional offset of the patient's initial lying center of gravity relative to the geometric center of the bed and the positional offset coefficient are calculated using a dynamic calculation model of center of gravity offset. Based on the patient's basic information and the bed's basic parameters, the overall restraint safety factor is calculated using the restraint tension safety factor algorithm. Based on the patient's basic information and the bed's basic parameters, the equipment's balance reserve coefficient is calculated using a balance reserve torque calculation model. The final safe lifting height is generated based on the body position offset coefficient, the overall restraint safety factor, and the equipment balance reserve factor. A synchronous lifting control signal sending module is used to send synchronous lifting control signals to the electric telescopic poles in each area to control the synchronous lifting of the electric telescopic poles in each area. The locking module is used to obtain the positioning signal from the electric telescopic rods in each area, generate a locking control signal to the electric telescopic rod in the middle area, and lock it as the rotation center axis. An initial control quantity generation module is used to calculate, based on a center of gravity balance model, the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head region, and to calculate the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot region. The calculation of the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head region based on the center of gravity balance model includes: Based on the patient's basic information and the bed's basic parameters, the reference center of gravity parameters of the electric telescopic rod in the head region are calculated using the head-cervical spine coupling center of gravity correction algorithm. Based on the reference center of gravity parameters of the electric telescopic pole in the head area, the relationship between the rotation angle and the telescopic amount of the electric telescopic pole in the head area is calculated by the rotation-telescopic coordinated phase matching algorithm. The initial deflection angle of the head, the real-time extension and retraction amount of the head, and the motion acceleration curve are generated based on the relationship between the rotation angle and extension amount of the electric telescopic rod in the head area. The calculation of the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area includes: Obtain a trained three-input neural network model of the foot; Foot features are generated based on the patient's basic information and the bed's basic parameters. Foot features are input into the foot three-input neural network model to obtain predicted values ​​for rotation angle, extension and contraction, motion acceleration, and bed board collaborative compensation coefficient. The predicted values ​​of rotation angle, extension and retraction, motion acceleration, and bed board coordination compensation coefficient are corrected to obtain the corrected predicted values ​​of rotation angle, extension and retraction, and motion acceleration. Motion acceleration curves are generated based on the corrected predicted values ​​of motion acceleration; among them, the corrected predicted value of rotation angle is used as the initial deflection angle of the foot, and the corrected predicted value of extension and contraction is used as the real-time extension and contraction of the foot. The foot-based three-input neural network model includes: The input layer contains 18 neurons; The number of hidden layers is 3. The output layer contains four neurons; wherein, The 18 neurons are used to receive the following features: Lower limb mass, joint angles between connecting plate 1 and connecting plate 2, joint angles between connecting plate 2 and connecting plate 3, joint angles between connecting plate 3 and connecting plate 4, joint torques between connecting plate 1 and connecting plate 2, joint torques between connecting plate 2 and connecting plate 3, joint torques between connecting plate 3 and connecting plate 4, moving state of movable bed board, moving speed of movable bed board, contact tension of thigh side fixing strap 1 and thigh fixing strap 2, contact tension of calf side fixing strap 1 and calf side fixing strap 2, fixing tension of elastic band on footrest, leg pressure sensor array data, leg blood flow velocity, distance from the long side foot area of ​​the rotating frame to the rotation center axis, maximum load of electric telescopic rod, dynamic deformation coefficient of footrest elastic band, joint damping coefficient of connecting plate assembly; The hidden layer has 64 neurons in the first layer, 32 neurons in the second layer, and 16 neurons in the third layer. A head region control signal generation module is used to generate a coordinated control signal for head deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's head region. The generation of the coordinated control signal includes: Obtain the dynamic deformation of the headrest and the cervical spine safety correction coefficient at the current moment; Based on the dynamic deformation of the head pad at the current moment and the cervical spine safety correction coefficient, the initial head deflection angle, the real-time head extension and contraction amount, and the head motion acceleration are corrected respectively, so as to obtain the corrected head target deflection angle, the corrected head motion acceleration, and the corrected head target extension and contraction amount. A head deflection and extension coordinated control signal is generated based on the corrected head target deflection angle, the corrected head motion acceleration, and the corrected head target extension and retraction. A foot area control signal generation module is used to generate a coordinated control signal for foot deflection and extension based on the initial deflection angle, real-time extension amount, and motion acceleration curve of the electric telescopic rod in the patient's foot area. The generation of the coordinated control signal includes: Obtain a trained, uncorrected neural network for the foot; Based on the patient's basic information, the bed's basic parameters, and the initial deflection angle, real-time extension and retraction amount, and motion acceleration curve of the electric telescopic rod in the foot area, a fusion feature is generated. The fused features are input into a trained foot-uncorrected neural network to obtain a coordinated control signal for foot deflection and extension. The foot-uncorrected neural network includes: The input layer includes 19 signal-specific feature neurons and 3 motion parameter reference value input terminals; A feature adaptation layer is used to generate a 32-dimensional fused feature vector based on the 22-dimensional standardized feature vector of the input layer. A signal conversion layer is used to generate a 24-dimensional signal adaptation feature vector based on the 32-dimensional fused feature vector output by the feature adaptation layer. The output layer is used to generate servo drive pulse width, linear motor drive voltage and speed closed-loop control voltage based on 24-dimensional signal adaptation feature vectors. A control signal transmitting module is used to transmit a head deflection and extension coordination control signal to the electric telescopic rod in the head area, so that the electric telescopic rod in the head area works according to the head deflection and extension coordination control signal, and to transmit a foot deflection and extension coordination control signal to the electric telescopic rod in the foot area, so that the electric telescopic rod in the foot area works according to the foot deflection and extension coordination control signal.

Images