Cusp targeting magnetic stimulation closed-loop control method and system based on flexible pressure array

Abstract

The application discloses a bony prominence targeting magnetic stimulation closed-loop control method and system based on a flexible pressure array, and belongs to the technical field of intelligent nursing and rehabilitation physiotherapy. The method collects human body bed pressure distribution data through a flexible pressure film array and generates pressure feature data, inputs a multi-channel feature map into a human body posture reconstruction model after extracting an effective contact area, identifies a lying posture type and a bony landmark candidate area, and then identifies a bony prominence compression area and calculates a bony prominence risk value; a spatial mapping relationship between the flexible pressure array and a magnetic stimulation coil array is established, corresponding coils are driven to output magnetic stimulation, and closed-loop feedback adjustment is performed according to pressure changes before and after stimulation. The application realizes a complete process from pressure perception, posture reconstruction, bony prominence identification to targeted magnetic stimulation output and closed-loop feedback, and improves the precision, safety and automation level of pressure injury risk identification and nursing intervention.

Description

Technical Field

[0001] This invention relates to the field of intelligent nursing and rehabilitation physiotherapy technology, and in particular to a closed-loop control method and system for targeted magnetic stimulation of bony prominences based on a flexible pressure array. Background Technology

[0002] Pressure injuries, also known as pressure ulcers or bedsores, are localized injuries to the skin and / or subcutaneous tissue caused by pressure or pressure combined with shear force, typically located at bony prominences. Pressure injuries are a key monitoring indicator for nursing quality and patient safety, and are listed as one of the five most common factors causing patient injury. Epidemiological surveys show that the incidence of pressure injuries among hospitalized patients in my country is 3.2%–5.2%. Pressure injuries mainly occur in individuals who are bedridden for extended periods, have limited mobility, sensory impairments, are undergoing postoperative rehabilitation, are in intensive care, or are receiving elderly care, and are particularly prone to occur in bony prominences such as the sacrum, ischial tuberosity, greater trochanter of the femur, scapula, elbow, and heel. The prevention and treatment of pressure injuries has become a major challenge in clinical nursing.

[0003] Currently, the main methods for preventing and intervening in pressure injuries include postural changes and manual care, support surface decompression devices, and electrical and magnetic stimulation therapy.

[0004] In terms of position changes and manual care, in clinical practice, nurses reduce the continuous pressure on local tissues by turning patients over at regular intervals. However, implementing regular position changes in busy nursing practice is challenging, and the quality of position changes varies. The frequency of turning patients needs to be individualized. This nursing model that relies on manual operation is not only labor-intensive, but also makes it difficult to ensure the continuous stability of the frequency and quality of turning.

[0005] Regarding pressure relief devices on the support surface, existing pressure ulcer mattresses mainly achieve passive pressure relief through methods such as alternating inflation and deflation of airbags and dispersing pressure with gel or sponge pads. Some products further integrate pressure-sensing mattresses to monitor the bed resting status and local pressure conditions, and assist caregivers in reducing the risk of pressure injuries through functions such as pressure threshold alarms, body position recognition, and bed exit detection. However, most pressure ulcer mattresses mainly achieve passive pressure relief through alternating airbag support or mechanical lifting, which cannot provide precise targeted intervention based on the actual contours of the human body, changes in posture, and the location of bony landmarks.

[0006] In the field of electrostimulation and magnetic stimulation therapy, some products combine electrostimulation or magnetic stimulation to treat pressure areas. For example, existing technology discloses a smart electrotherapy mattress with several flexible electrodes set in the concave areas of a standard human body contour. The corresponding electrodes are selected for electrotherapy by a controller. Alternating magnetic fields have also been studied to improve local blood circulation. Studies have shown that exposure to a 50Hz alternating magnetic field can quickly restore the blood flow to the soles of the feet that has decreased due to compression and fixation, and the effect can last for 40 minutes after the exposure ends. Magnetic fields have advantages such as being non-invasive and having a large tissue penetration depth, showing potential in improving microcirculation in ischemic tissues. However, existing magnetic therapy mattresses mostly use fixed magnetic areas or manually selected sites, and there is a lack of real-time matching between the stimulation location and the actual lying posture and changes in body position.

[0007] Existing pressure ulcer prevention mattresses and physiotherapy devices still have significant shortcomings. For example, patent ZL201920316996.5 discloses a pressure ulcer prevention pad with a pressure alarm function. It houses a controller, battery, A / D converter, storage device, wireless communication unit, and alarm within its housing, and a switch button on the outside. The pad itself includes a fabric layer and a filling layer. This solution monitors the pressure state using a pressure sensor. When the pressure exceeds a threshold and persists for a set time, an audible alarm is triggered to alert caregivers to assist with turning the patient. However, this solution is limited to pressure monitoring and audible alarms, only reminding caregivers to turn the patient. It does not establish a spatial mapping relationship between pressure data and the body's contour posture, bony prominences, and the actuation site. It lacks an active intervention mechanism and cannot automatically output differentiated intervention parameters based on different bony prominence locations.

[0008] Patent ZL202011199819.7 discloses a real-time dynamic prevention and detection system for pressure injuries based on pressure sensors. This system acquires pressure change data caused by bedridden patients and combines it with temperature and humidity changes at the pressure site. It then performs predictive analysis based on dynamically converted pressure injury detection rules to help assess the risk of pressure injuries. This solution focuses on risk detection and predictive analysis, primarily addressing the monitoring of pressure conditions and risk assessment. However, it does not establish a spatial mapping relationship between pressure data and human body contours, posture, bony prominences, and the actuation site. Furthermore, it lacks an active intervention mechanism and cannot automatically output differentiated magnetic stimulation parameters based on different bony prominence locations.

[0009] Patent application CN115054458A discloses a smart medical mattress for preventing pressure injuries and a control method therein. The mattress incorporates multiple electromagnetic structures. By using a magnetic sheet in conjunction with these structures, magnetic levitation is achieved in bony prominences such as the sacrum, scapula, and heels, temporarily suspending pressure on the skin and reducing localized pressure. However, this method relies on magnetic levitation for localized decompression, focusing on controlling support or levitation force rather than improving the condition of pressure-bearing tissues through magnetic stimulation. Furthermore, the bony prominence areas are primarily based on common supine positions and lack features such as flexible pressure array-based posture reconstruction, dynamic identification of bony prominence positions, and adaptive adjustment of coil stimulation parameters.

[0010] Patent ZL202011439200.9 discloses a functional electrical stimulation pressure ulcer prevention and control system with intelligent pressure feedback control. It includes a mobile intelligent terminal, a pressure ulcer prevention and control device, electrode pads, and a flexible thin-film pressure sensor. Different stimulation modes with different parameters are selected based on the magnitude and duration of pressure on the buttocks, and closed-loop control is achieved using body surface pressure as a feedback signal. However, this system uses a localized flexible thin-film pressure sensor, which does not form a high-density pressure array covering the bed surface, making it impossible to reconstruct the human body's contour posture and locate bony prominences. Its electrical stimulation requires electrode pads to contact the body surface, with fixed positions and manual placement, making it difficult to dynamically adjust the stimulation target point according to changes in body posture. Furthermore, its feedback signal is based solely on the magnitude and duration of local pressure, without establishing a bony prominence risk value model weighted by multiple factors such as bony prominence type, local pressure level, pressure area, position confidence, and postural stability, thus failing to achieve differentiated and precise intervention for different bony prominence areas.

[0011] In summary, existing pressure ulcer care mattresses and physiotherapy devices mainly rely on pressure threshold alarms, airbag decompression, mechanical support, or fixed-position electrical stimulation. These methods struggle to accurately identify and target pressure areas on bony prominences based on actual changes in the body's lying posture. Different bony prominences vary in tissue thickness, pressure intensity, tolerance, and intervention requirements. Current solutions typically fail to automatically match differentiated stimulation intensity, frequency, pulse width, duration, and safety limits for different bony joint locations such as the sacrum, hip, heel, and scapula, making it difficult to achieve individualized, dynamic, and precise intervention for pressure injury risks.

[0012] Therefore, there is an urgent need for a closed-loop control method and system that can collect bed pressure distribution in real time based on a flexible pressure array, reconstruct human body contour posture and locate the pressure area of ​​bony prominences by combining a deep learning model, and then drive the corresponding magnetic induction coil to implement targeted magnetic stimulation. Summary of the Invention

[0013] To achieve the above-mentioned objectives and other advantages of the present invention, a first objective of the present invention is to provide a closed-loop control method for bone prominence targeted magnetic stimulation based on a flexible pressure array, comprising the following steps: The pressure distribution data of the human body in bed state is continuously collected by a flexible pressure film array laid on the bed surface, and the collected pressure distribution data is preprocessed to generate pressure characteristic data. Based on the pressure characteristic data, the effective contact area between the human body and the bed surface is extracted to obtain the overall pressure profile of the human body on the bed surface, and a human body contact area map is generated based on the effective contact area. Based on the pressure feature data, a pressure distribution map, a pressure gradient map, and a historical pressure change sequence are generated. The pressure distribution map, the human body contact area map, the pressure gradient map, and the historical pressure change sequence are used as multi-channel feature maps and input into the human posture reconstruction model to reconstruct the human body contour posture and identify the lying posture type and candidate areas of bony landmarks. Based on the human body contour posture reconstruction results and the current pressure distribution, the pressure areas of bony prominences under different postures are identified, and the bony prominence risk value used to characterize the risk of pressure injury is calculated. Establish a spatial mapping relationship between the flexible pressure array and the magnetic stimulation coil array, map the identified bony prominence pressure area to the corresponding magnetic induction coil unit or coil combination, and generate control commands containing target coil position information; Based on the type of bony prominence region and the risk value of the bony prominence, corresponding magnetic stimulation parameters are generated. Based on the generated magnetic stimulation parameters and the control command, the magnetic induction coil at the corresponding position is driven to output magnetic stimulation; Pressure data continues to be collected after magnetic stimulation output, and closed-loop feedback adjustment is performed based on pressure changes and bony prominence risk changes before and after stimulation.

[0014] Furthermore, the flexible pressure film array is a 32×64 flexible pressure film pad, forming a pressure array consisting of 2048 pressure sampling points; The preprocessing includes zero-point calibration, outlier removal, smoothing filtering, and normalization. The pressure characteristic data includes the current pressure matrix, historical pressure sequence, pressure change, pressure center location, contact area, local high pressure points, and pressure gradient information.

[0015] Furthermore, the extraction of the effective contact area between the human body and the bed surface specifically includes: Areas with pressure values ​​below the contact threshold are classified as non-contact areas; Areas with pressure values ​​higher than the contact threshold and that are interconnected are identified as human contact areas. Connectivity analysis is performed on the human contact area to remove isolated noise points and non-human contact areas.

[0016] Furthermore, the human posture reconstruction model is a multi-task deep learning model, whose inputs include a pressure distribution map, a human contact area map, a pressure gradient map, and a historical pressure change sequence; its outputs include human body contour, human body axis direction, lying posture type, and candidate regions of bony landmarks. The types of lying positions include at least supine, left lateral, right lateral, semi-sitting, sitting up, sitting at the bedside, and getting out of bed.

[0017] Furthermore, the identification of the pressure area of ​​the bony prominence under different postures includes: When the lying position is supine, the scapula, sacrum, elbow and heel are the target areas for identifying the pressure points on bony prominences. When the lying position is left lateral or right lateral, the acromion region, the greater trochanter of the femur in the hip, the lateral side of the knee, and the ankle are used as the identification targets for the pressure areas of bony prominences. When the lying position is semi-recumbent or bedside sitting, the sacrococcygeal region, ischial tuberosity and heel are used as the identification targets for the pressure area of ​​bony prominences. For each identified target, the local pressure level, pressure area, duration of pressure, and location confidence are calculated. When the threshold conditions are met simultaneously, it is identified as a high-risk pressure area for bony prominences.

[0018] Furthermore, the risk value of the bony prominence is calculated as follows: Bone protrusion risk value = pressure intensity weight × local pressure level + time weight × duration of continuous pressure + location weight × bone protrusion location reliability + posture weight × posture stability; When the risk value of the bony prominence exceeds the preset risk threshold, the area of ​​the bony prominence under pressure is determined to meet the conditions for magnetic stimulation output.

[0019] Furthermore, the spatial mapping relationship is established in the following way: In the mattress structure, a flexible array pressure layer and a flexible PCB magnetic induction coil layer are sequentially arranged under the human body; Spatial registration is completed between the two layers to establish the correspondence between each pressure sampling point or pressure region and one or more adjacent magnetic induction coils; For the identified high-risk pressure areas of bony prominences, multiple magnetic induction coils near the target location are selected to work together to form a local magnetic stimulation area covering the pressure area of ​​the bony prominence.

[0020] Furthermore, the magnetic stimulation parameters include stimulation location, number of activation coils, stimulation intensity, stimulation frequency, pulse width, stimulation duration, and stimulation interval; Different initial coil configuration rules correspond to different bony prominence areas: the sacrococcygeal and ischial tuberosity areas use coil groups including multiple adjacent coils, the heel area uses one or two coils with a stimulation intensity lower than that of the sacrococcygeal and ischial tuberosity areas, and the scapula and hip areas select coil groups on the same side or opposite side according to the human posture.

[0021] Furthermore, when multiple bony prominence regions simultaneously meet the magnetic stimulation output conditions, the output priority is determined according to the risk value from high to low. When the number of drive channels is greater than or equal to the number of output areas, multiple areas are output in a time-sharing or partitioned manner. When the number of drive channels is less than the number of areas to be output, the area with the highest risk value is output first. During coil activation, coil temperature, drive current, operating time, and human body movement status are continuously monitored.

[0022] Furthermore, the closed-loop feedback adjustment includes: When the pressure risk value of the target area decreases after output, maintain the current parameters or gradually reduce the stimulation intensity; If the pressure risk value of the target area does not decrease after output and the safety limit is not triggered, adjust the stimulation parameters or expand the coverage of adjacent coils within the safe range. The magnetic stimulation output will automatically stop or pause when the system detects that a person is getting out of bed, turning over or sitting up, the coil temperature exceeds the safety threshold, the drive current is abnormal, the sensor data is abnormal, or the caregiver manually stops the stimulation.

[0023] Furthermore, it also includes the following steps: Nursing logs are generated by recording human posture, high-risk areas of bony prominences, pressure risk values, activation coil locations, magnetic stimulation parameters, output duration, safety status, and closed-loop adjustment results. The nursing terminal displays the current human lying position, pressure heat map, high-risk areas of bony prominences, the location of the working magnetic stimulation coil, and the risk change trend; When the risk value of the bony prominence continuously exceeds the preset risk threshold for more than the first preset time, the risk value does not decrease after intervention, the human body is not turned over for more than the second preset time, or the equipment is malfunctioning, a prompt message is sent to the nursing staff.

[0024] A second objective of this invention is to provide a closed-loop control system for bone protuberance-targeted magnetic stimulation based on a flexible pressure array, for performing the above-described method, the system comprising: A flexible pressure array sensing module is laid on the bed surface to continuously collect pressure distribution data of the human body while in bed. The data preprocessing module is used to perform zero-point calibration, outlier removal, smoothing filtering, and normalization on the collected pressure data to generate pressure characteristic data. The human body contact area extraction module is used to extract the effective contact area between the human body and the bed surface based on the pressure feature data, obtain the overall pressure profile of the human body on the bed surface, and generate a human body contact area map based on the effective contact area. The human posture reconstruction module is used to generate a pressure distribution map, a pressure gradient map, and a historical pressure change sequence based on the pressure feature data. The pressure distribution map, human contact area map, pressure gradient map, and historical pressure change sequence are input into a multi-task deep learning model to reconstruct the human contour posture and identify the lying posture type and candidate regions of bony landmarks. The bony prominence pressure area identification module is used to identify the bony prominence pressure area and calculate the bony prominence risk value to characterize the risk of pressure injury based on the posture reconstruction results and the current pressure distribution. The spatial mapping module is used to establish the correspondence between the pressure array coordinates and the magnetic stimulation coil coordinates, and to map the identified bony prominence pressure area to the corresponding magnetic induction coil unit or coil combination, generating control commands containing target coil position information; The magnetic stimulation parameter generation module is used to automatically generate magnetic stimulation parameters based on the type of bony prominence region and the risk value of the bony prominence. A flexible PCB magnetic induction coil array module is installed under the human body in a mattress. It includes multiple independently controllable magnetic induction coil units for outputting magnetic stimulation according to the control commands and magnetic stimulation parameters. The drive control module is used to drive the coil unit at the corresponding position to work, and to monitor the coil temperature, drive current and working time; The closed-loop feedback control module is used to dynamically adjust based on changes in pressure and bony prominence risk before and after stimulation, and to automatically stop or pause the magnetic stimulation output when an abnormal state is detected.

[0025] A third objective of the present invention is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0026] A fourth objective of the present invention is to provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0027] Compared with the prior art, the beneficial effects of the present invention are: This invention improves the accuracy and adaptability of identifying pressure zones on bony prominences. By continuously acquiring pressure distribution data of the human body in bed using a 32×64 flexible pressure array and combining it with a human contour posture reconstruction model, the system comprehensively identifies the body axis direction, lying posture type, and candidate bony landmarks. This allows the system to move beyond relying solely on single-point pressure peaks or fixed body positions to determine pressure risk. Instead, it dynamically determines the location of pressure-prone areas such as the sacrum, ischial tuberosities, hips, scapulae, elbows, and heels based on different postures such as supine, lateral, semi-recumbent, and sitting-up positions, thereby improving the accuracy of locating high-risk areas of bony prominences during prolonged bed rest.

[0028] This invention enables precise matching of magnetic stimulation locations with actual pressure areas on the human body. By pre-establishing a spatial correspondence between the flexible array pressure layer and the flexible PCB magnetic induction coil layer, the identified high-risk pressure areas of bony prominences are automatically mapped to the corresponding magnetic induction coil units or coil combinations. This allows magnetic stimulation to move dynamically with changes in body posture and the migration of pressure areas, improving the targeting of magnetic stimulation intervention and reducing ineffective stimulation and stimulation of non-target areas.

[0029] This invention improves the automation, safety, and individualization of nursing interventions. The system automatically matches stimulation intensity, frequency, duration, and coil combination based on the type of bony prominence, local pressure level, duration of pressure application, postural stability, and safety status. During stimulation, it continuously monitors pressure changes, body movement, bed ambulation status, coil temperature rise, and drive current. When abnormal body movement, bed ambulation, temperature rise, or current occurs, the system automatically pauses or stops output, thus forming a closed-loop control process from pressure monitoring, risk identification, magnetic stimulation execution to feedback adjustment, improving the real-time performance, reliability, and safety of nursing interventions for pressure injury risks.

[0030] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it according to the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Specific embodiments of the present invention are given in detail below with reference to the accompanying drawings. Attached Figure Description

[0031] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 Flowchart of a closed-loop control method for bone protuberance-targeted magnetic stimulation based on a flexible pressure array; Figure 2 This is a schematic diagram of the closed-loop control principle of bone protuberance targeted magnetic stimulation based on a flexible pressure array. Figure 3 Flowchart for extracting the effective contact area between the human body and the bed surface; Figure 4 Flowchart for identifying the pressure zones of bony prominences in different postures; Figure 5 Create a flowchart for spatial mapping relationships; Figure 6 This is a flowchart of the closed-loop feedback adjustment process; Figure 7 This is a schematic diagram of a computer device. Figure 8 This is a schematic diagram of a computer-readable storage medium. Detailed Implementation

[0032] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be noted that, without conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments.

[0033] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention.

[0034] The drawing numbers in this application are only used to distinguish the steps in the scheme and are not used to limit the execution order of the steps. The specific execution order is as described in the specification.

[0035] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0036] This invention provides a method and system for intervening in the risk of pressure injury based on a flexible pressure sensing array, human posture reconstruction, bony prominence region identification, and closed-loop magnetic stimulation control. This invention is applicable to scenarios such as long-term bed rest, limited mobility, postoperative rehabilitation, elderly care, and intensive care. It collects pressure distribution data of the human body in bed using a flexible pressure array, combines this with a deep learning model to infer the human body's contour posture and pressure-affected areas of bony landmarks, and links corresponding magnetic induction coils to output personalized stimulation parameters, achieving dynamic identification, targeted intervention, and safe closed-loop control of pressure-prone areas. The specific scheme is as follows: Example 1 A closed-loop control method for targeted magnetic stimulation of bony prominences based on a flexible pressure array, such as... Figure 1 , Figure 2 As shown, it includes the following steps: S100: Continuously collect pressure distribution data of the human body in bed state through a flexible pressure film array laid on the bed surface, and preprocess the collected pressure distribution data to generate pressure characteristic data. In this embodiment, a 32×64 flexible pressure film pad laid on the bed surface continuously collects pressure distribution data of the human body while in bed. The flexible pressure film pad forms a pressure array consisting of 2048 pressure sampling points. After the human body lies down, the flexible pressure film pad continuously collects pressure distribution data between the human body and the bed surface, obtaining a two-dimensional pressure matrix that changes over time.

[0037] It should be noted that the 32×64 specification (2048 sampling points) of the flexible pressure film array is the result of optimization based on the average human body surface area and bed size, which ensures both spatial coverage density of various bony prominence areas of the human body and controls the amount of data and hardware costs.

[0038] The system preprocesses the collected pressure data, including zero-point calibration, outlier removal, smoothing filtering, and normalization, to reduce the impact of sensor drift, local noise, and environmental interference on subsequent identification results.

[0039] Specifically, zero-point calibration involves collecting the baseline value of the pressure array when the person is not in bed, and subtracting the baseline value from the subsequent collected pressure values ​​to eliminate the initial bias of the sensor; outlier removal involves identifying sampling points whose pressure values ​​significantly deviate from the neighborhood mean, and using median filtering or methods based on neighborhood standard deviation to remove and interpolate for repair; smoothing filtering uses Gaussian filtering or mean filtering to smooth the pressure data and suppress high-frequency noise; normalization maps the pressure values ​​to a uniform data range, which is convenient for subsequent processing by deep learning models.

[0040] The preprocessed pressure data forms pressure feature data, including the current pressure matrix, historical pressure sequence, pressure change, pressure center location, contact area, local high-pressure points, and pressure gradient information. Specifically, the current pressure matrix is ​​the two-dimensional pressure distribution at the current moment; the historical pressure sequence is a sequence of pressure matrices from the past N moments, used to characterize pressure change trends; the pressure change is the difference between pressure matrices at adjacent moments; the pressure center location is the geometric center of the contact area between the human body and the bed surface; the contact area is the number of sampling points with pressure values ​​exceeding the contact threshold multiplied by the area per unit sampling point; local high-pressure points are the locations of sampling points with pressure values ​​exceeding the local neighborhood mean; and the pressure gradient information is the spatial gradient of the pressure matrix.

[0041] S200. Extract the effective contact area between the human body and the bed surface based on the pressure feature data to obtain the overall pressure profile of the human body on the bed surface, and generate a human body contact area map based on the effective contact area. The system extracts the effective contact area between the human body and the bed surface from the pressure matrix based on a pressure threshold. In this embodiment, as shown... Figure 3 As shown, the effective contact area between the human body and the bed surface specifically includes: S210. Areas with pressure values ​​below the contact threshold are determined to be non-contact areas; S220. Areas with pressure values ​​higher than the contact threshold and that are interconnected are identified as human contact areas. S230. Perform connected component analysis on the human contact area to remove isolated noise points and non-human contact areas. Specifically, the connected component analysis method is as follows: use a 4-connected or 8-connected algorithm to identify all connected components in the pressure matrix that are above the contact threshold, calculate the area of ​​each connected component (i.e., the number of sampling points it contains), and determine connected components with areas smaller than a preset area threshold as isolated noise points or small non-human contact areas and remove them. Retain the connected component with the largest area or multiple connected components with areas exceeding the preset threshold as the main human contact area.

[0042] It should be noted that the contact threshold setting needs to balance sensitivity and anti-interference capability. A contact threshold that is too low may cause noise to be misjudged as a contact area, while a contact threshold that is too high may cause low-pressure contact areas (such as the distal extremities) to be missed. In some specific implementations, the contact threshold is set to 3% to 8% of the full scale of the pressure array, and can be adaptively adjusted according to the actual mattress firmness and the patient's weight.

[0043] This step yields the overall pressure profile of the human body on the bed surface, which represents the projection range of the human body on the bed. Simultaneously, the system generates a human body contact area map based on the effective contact area. This map is a binary image, with contact areas marked as 1 and non-contact areas marked as 0, and is used as input for subsequent human posture reconstruction models.

[0044] S300. Generate a pressure distribution map, a pressure gradient map, and a historical pressure change sequence based on the pressure feature data. Input the pressure distribution map, human body contact area map, pressure gradient map, and historical pressure change sequence as multi-channel feature maps into the human posture reconstruction model to reconstruct the human body contour posture and identify the lying posture type and candidate areas of bony landmarks. The system inputs multi-channel feature maps generated from preprocessed, multi-frame pressure data into the human posture reconstruction model. These multi-channel feature maps include a pressure distribution map, a human contact area map, a pressure gradient map, and a historical pressure change sequence. The pressure distribution map is a visual representation of the current pressure matrix, and pressure values ​​can be converted to image pixel values ​​using color mapping or grayscale mapping. The human contact area map is a binary image, with pixel values ​​of 1 for contact areas and 0 for non-contact areas. The pressure gradient map is obtained by calculating the spatial gradient magnitude of the pressure matrix using the Sobel operator or difference operator. The historical pressure change sequence is a sequence of pressure changes over multiple recent moments, used to characterize the dynamic features of pressure evolution over time.

[0045] The human posture reconstruction model is a multi-task deep learning model. In some implementations, the model uses a convolutional neural network (CNN) as the backbone network, extracts spatial features of pressure distribution through multiple convolutional and pooling layers, extracts temporal features of historical pressure changes through a long short-term memory network (LSTM) or a gated recurrent unit (GRU), and finally outputs human contour, human body axis direction, lying posture type, and candidate regions of bony landmarks through multiple task branches.

[0046] Specifically, the human body contour is output using a semantic segmentation branch, marking the distribution range of various parts of the body on the pressure array coordinates. The body axis direction is output through a regression branch, showing the tilt angle and left-right rotation of the head and feet. The lying posture type is output through a classification branch, showing the current posture category, which includes at least supine, left lateral, right lateral, semi-recumbent, sitting up, bedside sitting, and off-bed state. Bony landmark candidate regions are output through object detection or keypoint detection branches, representing the possible locations of pressure-prone areas such as the sacrum, ischial tuberosity, greater trochanter of the femur, scapula, elbow, and heel on the pressure array coordinates.

[0047] Unlike methods that determine the pressure area solely based on the point of maximum pressure, this step comprehensively utilizes pressure pattern, pressure center, contact boundary, human symmetry, and temporal continuity to reconstruct the human body's posture in bed and the corresponding anatomical region distribution. For example, when the body is in a supine position, the pressure distribution exhibits a symmetrical trunk shape, with relatively flat morphology in the scapular and sacrococcygeal regions; when the body is in a lateral position, the pressure distribution exhibits an asymmetrical lateral shape, with concentrated pressure in the hip and acromion regions.

[0048] S400. Based on the human body contour posture reconstruction results and the current pressure distribution, identify the pressure-bearing areas of bony prominences under different postures, and calculate the bony prominence risk value used to characterize the risk of pressure injury. In some embodiments, such as Figure 4 As shown, the identification of the pressure area of ​​the bony prominence under different postures includes: S410. When the lying position is supine, the scapula, sacrum, elbow and heel are used as the identification targets for the pressure area of ​​bony prominences. For the supine position, the focus is on identifying the scapula, sacrococcygeal region, elbow, and heel. Specifically, based on the candidate location of the scapula in the candidate area of ​​bony landmarks, combined with the local pressure level and pressure area of ​​the corresponding location in the pressure distribution, it is determined whether the scapula is a pressure area of ​​the bony prominence; similarly, the sacrococcygeal region, elbow, and heel are determined separately.

[0049] S420. When the lying position is left lateral decubitus or right lateral decubitus, the acromion region, the greater trochanter of the femur in the hip, the lateral side of the knee, and the ankle are used as the identification targets for the pressure areas of bony prominences. For lateral decubitus positions (including left and right lateral decubitus), the focus is on identifying the acromion region, the greater trochanter of the femur in the hip, the lateral aspect of the knee, and the ankle. Specifically, based on the candidate location of the greater trochanter of the femur in the candidate bony landmark region, combined with the local pressure level at the corresponding location in the pressure distribution, it is determined whether the hip is a region of pressure on the bony prominence; similarly, the acromion region, the lateral aspect of the knee, and the ankle are determined separately.

[0050] S430. When the lying position is semi-recumbent or bedside sitting, the sacrococcygeal region, ischial tuberosity and heel are used as the identification targets for the bony prominence pressure area. For semi-reclining or bedside sitting positions, the focus is on identifying the sacrococcygeal region, ischial tuberosities, and heels. Specifically, the ischial tuberosities bear greater pressure in the semi-reclining position, and the determination is made based on the location of the ischial tuberosities in the candidate bony landmark areas, combined with the local pressure level at the corresponding location in the pressure distribution.

[0051] S440. For each identified target, calculate the local pressure level, pressure area, duration of pressure, and location confidence. When the threshold conditions are met simultaneously, it is determined as a high-risk pressure area for bony prominences.

[0052] Among them, the local pressure level is the average or peak pressure within the bony prominence region; the pressure area is the number of sampling points within the bony prominence region where the pressure value is higher than the contact threshold; the continuous pressure time is the length of time the bony prominence region is continuously pressured; and the location confidence is the location confidence of the bony prominence output by the human posture reconstruction model.

[0053] When a bony prominence region simultaneously meets the following conditions—local pressure level exceeding a threshold, duration of pressure exceeding a time threshold, and location confidence level meeting the requirements—the system identifies that region as a high-risk pressure region for the bony prominence. The system calculates the local pressure level, pressure area, duration of pressure, and location confidence level for each type of bony prominence region, and identifies it as a high-risk pressure region when all threshold conditions are met.

[0054] The system further calculates a risk value for bony prominences to characterize the risk of pressure injury. The calculation method for the risk value of bony prominences is as follows: Bone protrusion risk value = Pressure intensity weight × Local pressure level + Time weight × Duration of continuous pressure + Location weight × Bone protrusion positioning reliability + Posture weight × Postural stability The weights for pressure intensity, time, location, and posture can be preset based on clinical experience and experimental data. Postural stability is used to characterize the stability of the human body in the current posture; the more frequent the posture changes, the lower the postural stability. When the risk value of a bony prominence exceeds a preset risk threshold, the system determines that the pressure area of ​​that bony prominence requires magnetic stimulation intervention.

[0055] It should be noted that the weighting coefficients in the formula for calculating the risk value of bony prominences can be determined using the analytic hierarchy process (AHP) or regression analysis based on clinical data. In some specific implementations, the weight of pressure intensity is set to 0.35, the weight of time is set to 0.30, the weight of position is set to 0.20, and the weight of postural stability is set to 0.15. The preset risk threshold can be calibrated according to a clinical pressure injury risk assessment scale (such as the Braden scale).

[0056] S500: Establish a spatial mapping relationship between the flexible pressure array and the magnetic stimulation coil array, map the identified bony prominence pressure area to the corresponding magnetic induction coil unit or coil combination, and generate a control command containing target coil position information; In some embodiments, such as Figure 5 As shown, the spatial mapping relationship is established in the following way: S510. In the mattress structure, a flexible array pressure layer and a flexible PCB magnetic induction coil layer are sequentially arranged under the human body. S520: Spatial registration is completed between the two layers to establish the correspondence between each pressure sampling point or pressure region and one or more adjacent magnetic induction coils; The flexible array pressure layer is a 32×64 flexible pressure film pad, and the flexible PCB magnetic induction coil layer is a flexible PCB magnetic induction coil array. The two layers are set tightly together and spatially registered during production or assembly to ensure a precise correspondence between the coordinates of the pressure sampling points and the coordinates of the magnetic induction coils.

[0057] The system pre-establishes a correspondence between pressure sampling points and magnetic induction coils. Each pressure sampling point or pressure region can be mapped to one or more adjacent magnetic induction coils. Specifically, during spatial registration, the physical coordinates of each pressure sampling point on the mattress plane and the physical coordinates of each magnetic induction coil are recorded to establish a coordinate mapping table.

[0058] S530. For the identified high-risk pressure areas of bony prominences, select multiple magnetic induction coils near the target location to work together to form a local magnetic stimulation area covering the pressure area of ​​the bony prominence.

[0059] Specifically, for identified high-risk pressure areas of bony prominences, the system automatically locates the corresponding coil unit or coil combination in the flexible PCB magnetic induction coil layer based on its position in the pressure array. To avoid excessively small stimulation range or positional deviation, the system selects multiple small-sized magnetic induction coils near the target location to work together, forming a localized magnetic stimulation area covering the pressure area of ​​the bony prominence.

[0060] The system generates control commands containing the target coil's position information, which are used to drive the magnetic induction coil at the corresponding position to perform magnetic stimulation.

[0061] S600. Based on the type of bony prominence region and the risk value of the bony prominence, generate corresponding magnetic stimulation parameters; wherein, the magnetic stimulation parameters include stimulation location, number of activation coils, stimulation intensity, stimulation frequency, pulse width, stimulation duration and stimulation interval.

[0062] It should be noted that the safety range of magnetic stimulation parameters should comply with relevant medical device safety standards. In some specific implementations, the stimulation intensity should not exceed 50 mT, the stimulation frequency should not exceed 100 Hz, the duration of a single stimulation session should not exceed 40 minutes, and the stimulation interval should not be less than 2 hours. The safe threshold for coil temperature is set at 41℃, and the safe upper limit for drive current is determined according to the coil specifications.

[0063] Different bony prominence regions correspond to different initial parameter rules. Specifically, the sacrococcygeal and ischial tuberosities are located in the core of the human trunk, with thicker soft tissue coverage, requiring deeper magnetic field penetration and a larger coverage area. Therefore, coil groups including multiple adjacent coils are used to form a larger stimulation area and a stronger magnetic field intensity. The heel has thinner soft tissue coverage and denser nerve endings, requiring a smaller stimulation range and lower stimulation intensity to avoid discomfort. Therefore, single or two coils are used, with stimulation intensity lower than that of the sacrococcygeal and ischial tuberosities. The scapular and hip regions are located on both sides of the trunk, and coil groups on the same side or opposite side are selected according to the body posture; for example, in the lateral decubitus position, only the coil group located on the pressure side (lower side) is activated; in the supine position, both sides can be activated simultaneously.

[0064] In some specific implementations, the initial parameters for the sacrococcygeal and ischial tuberosity regions are: stimulation intensity 20-40 mT, stimulation frequency 20-50 Hz, pulse width 200-400 μs, and stimulation duration 15-30 minutes. The initial parameters for the heel are: stimulation intensity 5-15 mT, stimulation frequency 10-30 Hz, pulse width 100-200 μs, and stimulation duration 10-20 minutes. The initial parameters for the scapular and hip regions fall between these two values ​​and are dynamically adjusted according to the specific degree of pressure.

[0065] The parameter matching logic is as follows: the higher the risk of pressure, the higher the stimulation intensity or duration within the safe range; the longer the continuous pressure time, the higher the stimulation priority; when the posture changes significantly, the bony prominence position is re-identified and the coil is re-matched; when the human body leaves the bed, there is violent body movement, or the equipment temperature rises abnormally, the stimulation output is immediately reduced or stopped.

[0066] The system does not use fixed stimulation locations and parameters, but dynamically generates intervention plans based on current pressure risk and human posture, thereby achieving personalized and zoned magnetic stimulation intervention for different bony prominence areas.

[0067] S700. Based on the generated magnetic stimulation parameters and the control command, drive the magnetic induction coil at the corresponding position to output magnetic stimulation; In this embodiment, the control module drives the coil units at corresponding positions in the flexible PCB magnetic induction coil layer to work according to the generated magnetic stimulation parameters. Each coil unit can be controlled independently or can work collaboratively with adjacent coils to form a coil group.

[0068] When multiple bony prominences simultaneously meet the intervention criteria, the system determines the stimulation priority based on risk value, duration of pressure, and importance of the location. Specifically, the higher the risk value, the longer the duration of pressure, and the more important the location (e.g., the sacrococcygeal region is more important than the heel), the higher the stimulation priority.

[0069] If the number of driving channels is sufficient (i.e., the number of driving channels is greater than or equal to the number of output regions), multiple regions are stimulated in a time-sharing or zone-sharing manner. Time-sharing stimulation means that different regions are stimulated sequentially at different times; zone-sharing stimulation means that different regions are stimulated simultaneously with different parameters. If the number of driving channels is limited (i.e., the number of driving channels is less than the number of output regions), the region with the highest risk value is output first.

[0070] During coil activation, the system continuously monitors coil temperature, drive current, operating time, and human body movement to ensure the magnetic stimulation process remains within safe limits. If any monitored parameter exceeds a safe threshold, the system automatically reduces the stimulation intensity or pauses the stimulation output.

[0071] S800 continues to collect pressure data after magnetic stimulation output, and performs closed-loop feedback adjustment based on pressure changes and changes in bony prominence risk before and after stimulation.

[0072] Specifically, after magnetic stimulation is performed, the system continues to collect data from the flexible pressure array and compares changes in pressure, body posture, and bony prominence risk before and after stimulation. In some embodiments, such as Figure 6 As shown, the closed-loop feedback adjustment includes: S810. When the pressure risk value of the target area decreases after output, maintain the current parameters or gradually reduce the stimulation intensity. For example, if the risk value of the sacrococcygeal bony prominence decreases from high risk to medium risk, the system can gradually reduce the stimulation intensity from 30mT to 20mT, or shorten the stimulation duration from 20 minutes to 15 minutes.

[0073] S820: If the pressure risk value of the target area does not decrease after output and the safety limit is not triggered, the stimulation parameters can be adjusted or the coverage of adjacent coils can be expanded within the safe range. For example, if the risk value of the sacrococcygeal bony prominence does not change significantly after stimulation, the system can increase the stimulation intensity or increase the number of activated coils within the safe upper limit to enhance the effect of magnetic stimulation.

[0074] S830 automatically stops or pauses magnetic stimulation output when it detects that a person leaves the bed, the person is turning over or sitting up, the coil temperature exceeds the safety threshold, the drive current is abnormal, the sensor data is abnormal, or the caregiver manually stops it.

[0075] Through this closed-loop feedback mechanism, the system can dynamically adjust the magnetic stimulation strategy according to changes in the body's condition and stress risk, avoiding ineffective or excessive intervention caused by fixed parameter stimulation.

[0076] In some embodiments, the steps further include: The nursing log is generated by recording human posture, high-risk areas of bony prominences, pressure risk values, activation coil locations, magnetic stimulation parameters, output duration, safety status, and closed-loop adjustment results. The nursing log can be stored in chronological order to form a complete nursing record archive.

[0077] The nursing terminal displays the current patient's lying position, a pressure heat map, high-risk areas of bony prominences, the location of the active magnetic stimulation coil, and the trend of risk changes. Specifically, the nursing terminal can use a tablet computer, nursing station display screen, or mobile application to present the above information graphically. The pressure heat map uses color intensity to represent pressure levels, high-risk areas of bony prominences are highlighted with special markers (such as red circles or flashing indicators), and the location of the active magnetic stimulation coil is marked with a bright color on the mattress diagram.

[0078] When the risk value of a bony prominence continuously exceeds the preset risk threshold for more than the first preset time, the risk value does not decrease after intervention, the patient is not turned over for more than the second preset time, or there is equipment malfunction, a prompt message is sent to the nursing staff. The prompt message can be sent via sound alarm, pop-up notification, SMS push, or nursing station notification.

[0079] When the system identifies a persistently high-risk pressure area, when intervention does not provide significant improvement, when the patient has not been turned over for an extended period, or when the equipment malfunctions, it will alert the nursing staff so that they can take appropriate measures such as manual turning, repositioning, or further nursing interventions.

[0080] The overall workflow of this invention will be described below using a specific application scenario.

[0081] The patient is bedridden for an extended period on a bed equipped with the system of this invention. Upon system startup, a flexible pressure diaphragm array collects real-time pressure distribution data between the patient and the bed surface. The data preprocessing module performs zero-point calibration, outlier removal, smoothing filtering, and normalization on the collected pressure data to generate pressure characteristic data.

[0082] The human body contact area extraction module extracts the effective contact area between the patient and the bed surface based on pressure feature data, obtaining the overall pressure profile and human body contact area map. The human posture reconstruction module inputs the pressure distribution map, human body contact area map, pressure gradient map, and historical pressure change sequence into the deep learning model, and outputs the patient's current lying position type (e.g., supine) and candidate bony landmark regions.

[0083] The bony prominence pressure area identification module identifies the pressure on the scapula, sacrum, coccyx, elbow, and heel areas based on the supine position. Calculations show that bony prominences in the sacrum and coccyx have a risk value exceeding a preset risk threshold, classifying them as high-risk pressure areas.

[0084] The spatial mapping module maps the pressure area of ​​the sacrococcygeal region to the corresponding coil group in the flexible PCB magnetic induction coil layer, generating control commands containing the target coil position information. The magnetic stimulation parameter generation module matches and generates corresponding magnetic stimulation parameters based on the type and risk value of the sacrococcygeal bony prominence region, such as stimulation intensity of 30mT, stimulation frequency of 30Hz, and stimulation duration of 20 minutes.

[0085] The drive control module drives the coil group at the corresponding position to output magnetic stimulation according to the magnetic stimulation parameters and control commands. During the magnetic stimulation process, the closed-loop feedback control module continuously monitors the coil temperature and the patient's body movement.

[0086] Twenty minutes after the magnetic stimulation is applied, the closed-loop feedback control module determines whether the pressure risk value in the sacrococcygeal region has decreased based on the pressure data before and after stimulation. If it has decreased to a safe level, the current parameters are maintained or the stimulation intensity is gradually reduced; if it has not decreased and the safety limit has not been triggered, the stimulation parameters are adjusted for a new round of intervention.

[0087] The nursing terminal displays the patient's lying position, pressure thermogram, high-risk area markings in the sacrum and coccyx, and magnetic stimulation status in real time, and generates a nursing log. If the patient has not been turned over for more than 2 hours, the nursing terminal sends a turning prompt to the nursing staff.

[0088] Through the above process, the present invention realizes a complete automated nursing intervention loop from pressure sensing, posture reconstruction, bony prominence recognition to targeted magnetic stimulation output and closed-loop feedback.

[0089] It should be noted that the parameter ranges and thresholds described above are examples in specific implementations and should not be construed as limiting the scope of protection of this invention. Those skilled in the art can reasonably adjust the parameters according to actual application scenarios and clinical needs.

[0090] Example 2 Based on the same concept, this embodiment also provides a closed-loop control system for bone protrusion targeted magnetic stimulation based on a flexible pressure array, which applies the closed-loop control method for bone protrusion targeted magnetic stimulation based on a flexible pressure array provided in Embodiment 1. For a detailed description of the closed-loop control method for bone protrusion targeted magnetic stimulation based on a flexible pressure array provided in Embodiment 1, please refer to the corresponding description in Embodiment 1, which will not be repeated here.

[0091] It is understood that the bony prominence-targeted magnetic stimulation closed-loop control system based on a flexible pressure array provided in this embodiment includes corresponding hardware structures and / or software modules for executing each function in order to achieve the above-mentioned functions. Combining the units and algorithm steps of the various examples disclosed in this embodiment, this embodiment can be implemented in hardware or a combination of hardware and computer software. Whether a certain function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the technical solution of this embodiment.

[0092] A closed-loop control system for bone protuberance-targeted magnetic stimulation based on a flexible pressure array, such as Figure 1 , Figure 2 As shown, the system includes: A flexible pressure array sensing module, laid on the bed surface, is used to continuously collect pressure distribution data of the human body while in bed. In this embodiment, the flexible pressure array sensing module is a 32×64 flexible pressure film pad, forming a pressure array composed of 2048 pressure sampling points. It is laid on the bed surface, and after the human body lies down, it continuously collects the pressure distribution data between the human body and the bed surface, obtaining a two-dimensional pressure matrix that changes over time.

[0093] The data preprocessing module performs zero-point calibration, outlier removal, smoothing filtering, and normalization on the collected pressure data to generate pressure feature data. Zero-point calibration involves collecting the baseline value of the pressure array when the person is not in bed, and subtracting the baseline value from subsequent pressure values ​​to eliminate initial sensor bias. Outlier removal identifies sampling points whose pressure values ​​significantly deviate from the neighborhood mean, and uses median filtering or methods based on neighborhood standard deviation for removal and interpolation repair. Smoothing filtering uses Gaussian filtering or mean filtering to smooth the pressure data and suppress high-frequency noise. Normalization maps the pressure values ​​to a uniform data range, facilitating subsequent processing by deep learning models.

[0094] The preprocessed pressure data forms pressure feature data, including the current pressure matrix, historical pressure sequence, pressure change, pressure center location, contact area, local high pressure points, and pressure gradient information.

[0095] The human body contact area extraction module is used to extract the effective contact area between the human body and the bed surface based on the pressure feature data, obtain the overall pressure profile of the human body on the bed surface, and generate a human body contact area map based on the effective contact area.

[0096] Specifically, the human contact area extraction module classifies areas with pressure values ​​below the contact threshold as non-contact areas and areas with pressure values ​​above the contact threshold that are interconnected as human contact areas. It then performs connected component analysis on the human contact areas to remove isolated noise points and non-human contact areas. Specifically, the connected component analysis uses a 4-connected or 8-connected algorithm to identify all connected components in the pressure matrix that are above the contact threshold. It calculates the area of ​​each connected component, and identifies connected components with areas smaller than a preset area threshold as isolated noise points or non-human contact areas and removes them. The connected component with the largest area or multiple connected components with areas exceeding the preset threshold are retained as the main human contact areas.

[0097] The human posture reconstruction module is used to generate a pressure distribution map, a pressure gradient map, and a historical pressure change sequence based on the pressure feature data. The pressure distribution map, human contact area map, pressure gradient map, and historical pressure change sequence are input into a multi-task deep learning model to reconstruct the human contour posture and identify the lying posture type and candidate regions of bony landmarks.

[0098] Among them, the pressure distribution map is a visual representation of the current pressure matrix, and the pressure value can be converted into image pixel value through color mapping or grayscale mapping; the human contact area map is a binary image, with the pixel value of the contact area being 1 and the pixel value of the non-contact area being 0; the pressure gradient map is obtained by calculating the spatial gradient magnitude of the pressure matrix using the Sobel operator or the difference operator; the historical pressure change sequence is a sequence of pressure changes at multiple recent moments, used to characterize the dynamic characteristics of pressure evolution over time.

[0099] The human posture reconstruction model uses a convolutional neural network (CNN) as the backbone network. It extracts the spatial features of pressure distribution through multiple convolutional and pooling layers, and extracts the temporal features of historical pressure changes through a long short-term memory network (LSTM) or a gated recurrent unit (GRU). Finally, it outputs the human body contour, human body axis direction, lying posture type, and candidate regions of bony landmarks through multiple task branches.

[0100] The human body contour is output through a semantic segmentation branch, marking the distribution range of various parts of the human body on the pressure array coordinates; the human body axis direction is output through a regression branch, showing the tilt angle and left-right rotation of the head and feet; the lying posture type is output through a classification branch, showing the current posture category, which includes at least supine, left lateral, right lateral, semi-recumbent, sitting up, bedside sitting, and off-bed status; the candidate regions of bony landmarks are output through a target detection or key point detection branch, representing the possible locations of pressure-prone areas such as the sacrum, ischial tuberosity, greater trochanter of the femur, scapula, elbow, and heel in the pressure array coordinates.

[0101] The bony prominence pressure area identification module is used to identify the bony prominence pressure area based on the posture reconstruction results and the current pressure distribution, and to calculate the bony prominence risk value to characterize the risk of pressure injury.

[0102] Specifically, when the lying position is supine, the scapula, sacrococcygeal region, elbow, and heel are used as the identification targets for bony prominence compression areas; when the lying position is left or right lateral decubitus, the acromion region, greater trochanter of the femur, lateral knee, and ankle are used as the identification targets; when the lying position is semi-recumbent or bedside sitting, the sacrococcygeal region, ischial tuberosity, and heel are used as the identification targets. For each identification target, the local pressure level, pressure area, duration of pressure, and location confidence are calculated. When all threshold conditions are met, it is identified as a high-risk bony prominence compression area.

[0103] The risk value of the bony prominence is calculated as follows: Bony prominence risk value = pressure intensity weight × local pressure level + time weight × duration of continuous pressure + location weight × bony prominence location reliability + posture weight × posture stability.

[0104] The weights for pressure intensity, time, location, and posture can be preset based on clinical experience and experimental data. Postural stability is used to characterize the stability of the human body in the current posture; the more frequent the posture changes, the lower the postural stability. When the risk value of a bony prominence exceeds a preset risk threshold, the pressure area of ​​that bony prominence is deemed to meet the conditions for magnetic stimulation output.

[0105] The spatial mapping module is used to establish the correspondence between the coordinates of the pressure array and the coordinates of the magnetic stimulation coil, and to map the identified bony prominence pressure area to the corresponding magnetic induction coil unit or coil combination, generating control commands containing target coil position information.

[0106] Specifically, in the mattress structure, a flexible array pressure layer and a flexible PCB magnetic induction coil layer are sequentially arranged beneath the human body. Spatial registration is completed between the two layers to establish a correspondence between each pressure sampling point or pressure area and one or more adjacent magnetic induction coils. The flexible array pressure layer is a 32×64 flexible pressure film pad, and the flexible PCB magnetic induction coil layer is a flexible PCB magnetic induction coil array. The two layers are set tightly together and spatially registered during production or assembly to ensure a precise correspondence between the coordinates of the pressure sampling points and the coordinates of the magnetic induction coils. During the spatial registration process, the physical coordinates of each pressure sampling point on the mattress surface and the physical coordinates of each magnetic induction coil are recorded to establish a coordinate mapping table.

[0107] The magnetic stimulation parameter generation module is used to automatically generate magnetic stimulation parameters based on the type of bony prominence region and the risk value of the bony prominence. These parameters include stimulation location, number of activation coils, stimulation intensity, stimulation frequency, pulse width, stimulation duration, and stimulation interval.

[0108] Different initial parameter rules correspond to different bony prominence regions: the sacrococcygeal and ischial tuberosity regions use coil groups including multiple adjacent coils, the heel region uses one or two coils with a stimulation intensity lower than that of the sacrococcygeal and ischial tuberosity regions, and the scapula and hip regions use coil groups on the same side or opposite side depending on the human posture.

[0109] The parameter matching logic is as follows: the higher the risk of pressure, the higher the stimulation intensity or duration within the safe range; the longer the continuous pressure time, the higher the stimulation priority; when the posture changes significantly, the bony prominence position is re-identified and the coil is re-matched; when the human body leaves the bed, there is violent body movement, or the equipment temperature rises abnormally, the stimulation output is immediately reduced or stopped.

[0110] A flexible PCB magnetic induction coil array module, positioned beneath the body in the mattress, comprises multiple independently controllable magnetic induction coil units for outputting magnetic stimulation according to control commands and magnetic stimulation parameters. Each coil unit can be controlled independently or work collaboratively with adjacent coils to form a coil group. For identified high-risk pressure areas of bony prominences, the system automatically locates the corresponding coil unit or coil combination within the flexible PCB magnetic induction coil layer based on its position in the pressure array. Multiple small-sized magnetic induction coils near the target location are then selected to work together, forming a localized magnetic stimulation area covering the pressure area of ​​the bony prominence.

[0111] The drive control module is used to drive the coil unit at the corresponding position and monitor the coil temperature, drive current, and operating time. During coil activation, the drive control module continuously monitors the coil temperature, drive current, operating time, and the human body's movement status to ensure that the magnetic stimulation process remains within a safe range. When any monitored parameter exceeds the safety threshold, the drive control module automatically reduces the stimulation intensity or pauses the stimulation output.

[0112] The closed-loop feedback control module is used to dynamically adjust based on changes in pressure and bony prominence risk before and after stimulation, and to automatically stop or pause the magnetic stimulation output when an abnormal state is detected.

[0113] After magnetic stimulation is performed, the closed-loop feedback control module continues to collect data from the flexible pressure array and compares the changes in pressure, body posture, and bony prominence risk before and after stimulation. Specifically, if the pressure risk value in the target area decreases after output, the current parameters are maintained or the stimulation intensity is gradually reduced; if the pressure risk value in the target area does not decrease after output and safety limits are not triggered, the stimulation parameters are adjusted within a safe range or the coverage of adjacent coils is expanded; if the system detects that the body is getting out of bed, the body is significantly turning over or sitting up, the coil temperature exceeds the safety threshold, the drive current is abnormal, the sensor data is abnormal, or the caregiver manually stops the stimulation, the magnetic stimulation output is automatically stopped or paused.

[0114] The human-computer interaction and nursing record module displays the current patient's lying position, pressure heat map, high-risk areas of bony prominences, the location and risk trends of the active magnetic stimulation coil, generates a nursing log, and sends alerts to nursing staff when abnormal conditions are detected. The nursing terminal can be a tablet, nursing station display screen, or mobile application, presenting the above information graphically. The nursing log is stored chronologically, forming a complete nursing record archive.

[0115] It should be noted that the functional division of the modules in the above system embodiment is only an example. In actual implementation, the modules can be merged, split, or reorganized as needed, as long as the technical solution of the present invention can be achieved. The modules described in the present invention can be implemented by software, hardware, firmware, or any combination thereof.

[0116] This system uses a 32×64 flexible pressure film array laid on the bed surface as the sensing unit to continuously collect pressure distribution data of the human body in bed. After pressure calibration, noise suppression, contact area segmentation, and temporal feature extraction, the data is input into a human posture reconstruction model to obtain the human contour, lying posture type, body axis direction, and location of bony landmarks. The system further combines pressure peaks, pressure area, duration of pressure application, and posture changes to identify high-risk areas prone to pressure injuries, such as the sacrum, ischial tuberosity, hip, scapula, elbow, and heel, and maps these areas to the corresponding coordinates of the magnetic induction coil array built into the mattress. The control module automatically generates magnetic stimulation intensity, frequency, pulse width, duration, and working coil combinations based on the anatomical location, pressure level, and safety threshold of different bony prominence areas, driving the corresponding coils to implement targeted intervention. It also dynamically adjusts and safely stops based on pressure feedback, body movement, and coil temperature rise.

[0117] Through the above system, the present invention realizes a complete automated nursing intervention closed loop, from pressure sensing, posture reconstruction, bony prominence recognition to targeted magnetic stimulation output and closed-loop feedback.

[0118] Example 3 A computer device 900, such as Figure 7 As shown, the system includes a memory 910, a processor 920, and a computer program 930 stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a closed-loop control method for bony prominence-targeted magnetic stimulation based on a flexible pressure array. For a detailed description of the method, please refer to the corresponding description in the above method embodiments; it will not be repeated here.

[0119] Example 4 A computer-readable storage medium, such as Figure 8 As shown, a computer program is stored thereon. When executed by a processor, the computer program implements the steps of a closed-loop control method for bony prominence-targeted magnetic stimulation based on a flexible pressure array. For a detailed description of the method, please refer to the corresponding description in the above method embodiments, which will not be repeated here.

[0120] To address the problem that existing pressure ulcer care mattresses and physiotherapy devices mainly rely on pressure threshold alarms, airbag decompression, mechanical support, or fixed-position electrical stimulation, which make it difficult to accurately identify and target pressure areas of bony prominences based on actual changes in human lying posture, this invention provides a closed-loop intervention method and system for targeted magnetic stimulation of bony prominences based on flexible pressure array human posture reconstruction. This invention collects human pressure distribution data while in bed, reconstructs human contour posture using a deep learning model, dynamically identifies pressure-prone bony landmarks such as the sacrum, coccyx, hip, ischial tuberosity, scapula, elbow, and heel, and establishes a spatial mapping relationship between pressure array coordinates and magnetic stimulation coil coordinates. It automatically selects the corresponding coils and stimulation parameters to achieve personalized, zoned, and closed-loop magnetic stimulation intervention for different pressure areas of bony prominences, thereby improving the accuracy, safety, and effectiveness of pressure injury risk identification and nursing intervention.

[0121] The number of devices and processing scale described herein are for the purpose of simplifying the description of the invention. Applications, modifications, and variations of the invention will be readily apparent to those skilled in the art.

[0122] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

[0123] The apparatus, computer device, and non-volatile computer storage medium and method provided in the embodiments of this specification are corresponding. Therefore, the apparatus, computer device, and non-volatile computer storage medium also have similar beneficial technical effects as the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the corresponding apparatus, computer device, and non-volatile computer storage medium will not be repeated here.

[0124] Those skilled in the art will also know that, besides implementing the controller in the form of purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller take the form of logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices included within it for implementing various functions can also be considered structures within that hardware component. Alternatively, the devices for implementing various functions can be considered as both software units implementing the method and structures within a hardware component.

[0125] The systems, apparatuses, or units described in the above embodiments can be implemented by computer chips or physical entities, or by products with certain functions. For ease of description, the above apparatuses are described separately as various units based on their functions. Of course, when implementing one or more embodiments of this specification, the functions of each unit can be implemented in one or more software and / or hardware.

[0126] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0127] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0128] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0129] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0130] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0131] This specification may be described in the general context of computer-executable instructions, such as program units, that are executed by a computer. Generally, program units include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification may also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program units may reside in local and remote computer storage media, including storage devices.

[0132] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0133] The above description is merely an embodiment of this specification and is not intended to limit the scope of one or more embodiments of this specification. Various modifications and variations can be made to one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of one or more embodiments of this specification.

Claims

1. A closed-loop control method for bone protuberance-targeted magnetic stimulation based on a flexible pressure array, characterized in that, Includes the following steps: The pressure distribution data of the human body in bed state is continuously collected by a flexible pressure film array laid on the bed surface, and the collected pressure distribution data is preprocessed to generate pressure characteristic data. Based on the pressure characteristic data, the effective contact area between the human body and the bed surface is extracted to obtain the overall pressure profile of the human body on the bed surface, and a human body contact area map is generated based on the effective contact area. Based on the pressure feature data, a pressure distribution map, a pressure gradient map, and a historical pressure change sequence are generated. The pressure distribution map, the human body contact area map, the pressure gradient map, and the historical pressure change sequence are used as multi-channel feature maps and input into the human posture reconstruction model to reconstruct the human body contour posture and identify the lying posture type and candidate areas of bony landmarks. Based on the human body contour posture reconstruction results and the current pressure distribution, the pressure areas of bony prominences under different postures are identified, and the bony prominence risk value used to characterize the risk of pressure injury is calculated. Establish a spatial mapping relationship between the flexible pressure array and the magnetic stimulation coil array, map the identified bony prominence pressure area to the corresponding magnetic induction coil unit or coil combination, and generate control commands containing target coil position information; Based on the type of bony prominence region and the risk value of the bony prominence, corresponding magnetic stimulation parameters are generated. Based on the generated magnetic stimulation parameters and the control command, the magnetic induction coil at the corresponding position is driven to output magnetic stimulation; Pressure data continues to be collected after magnetic stimulation output, and closed-loop feedback adjustment is performed based on pressure changes and bony prominence risk changes before and after stimulation.

2. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, The flexible pressure film array is a 32×64 flexible pressure film pad, forming a pressure array consisting of 2048 pressure sampling points; The preprocessing includes zero-point calibration, outlier removal, smoothing filtering, and normalization. The pressure characteristic data includes the current pressure matrix, historical pressure sequence, pressure change, pressure center location, contact area, local high pressure points, and pressure gradient information.

3. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, The specific areas for extracting the effective contact area between the human body and the bed surface include: Areas with pressure values ​​below the contact threshold are classified as non-contact areas; Areas with pressure values ​​higher than the contact threshold and that are interconnected are identified as human contact areas. Connectivity analysis is performed on the human contact area to remove isolated noise points and non-human contact areas.

4. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, The human posture reconstruction model is a multi-task deep learning model. Its inputs include a pressure distribution map, a human contact area map, a pressure gradient map, and a historical pressure change sequence. Its outputs include human body contour, human body axis direction, lying posture type, and candidate regions of bony landmarks. The types of lying positions include at least supine, left lateral, right lateral, semi-sitting, sitting up, sitting at the bedside, and getting out of bed.

5. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 4, characterized in that, The identification of bony prominence compression areas under different postures includes: When the lying position is supine, the scapula, sacrum, elbow and heel are the target areas for identifying the pressure points on bony prominences. When the lying position is left lateral or right lateral, the acromion region, the greater trochanter of the femur in the hip, the lateral side of the knee, and the ankle are used as the identification targets for the pressure areas of bony prominences. When the lying position is semi-recumbent or bedside sitting, the sacrococcygeal region, ischial tuberosity and heel are used as the identification targets for the pressure area of ​​bony prominences. For each identified target, the local pressure level, pressure area, duration of pressure, and location confidence are calculated. When the threshold conditions are met simultaneously, it is identified as a high-risk pressure area for bony prominences.

6. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, The calculation method for the risk value of the bony prominence is as follows: Bone protrusion risk value = pressure intensity weight × local pressure level + time weight × duration of continuous pressure + location weight × bone protrusion location reliability + posture weight × posture stability; When the risk value of the bony prominence exceeds the preset risk threshold, the area of ​​the bony prominence under pressure is determined to meet the conditions for magnetic stimulation output.

7. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, The spatial mapping relationship is established in the following way: In the mattress structure, a flexible array pressure layer and a flexible PCB magnetic induction coil layer are sequentially arranged under the human body; Spatial registration is completed between the two layers to establish the correspondence between each pressure sampling point or pressure region and one or more adjacent magnetic induction coils; For the identified high-risk pressure areas of bony prominences, multiple magnetic induction coils near the target location are selected to work together to form a local magnetic stimulation area covering the pressure area of ​​the bony prominence.

8. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, The magnetic stimulation parameters include stimulation location, number of activation coils, stimulation intensity, stimulation frequency, pulse width, stimulation duration, and stimulation interval. Different initial coil configuration rules correspond to different bony prominence areas: the sacrococcygeal and ischial tuberosity areas use coil groups including multiple adjacent coils, the heel area uses one or two coils with a stimulation intensity lower than that of the sacrococcygeal and ischial tuberosity areas, and the scapula and hip areas select coil groups on the same side or opposite side according to the human posture.

9. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, When multiple bony prominence regions simultaneously meet the magnetic stimulation output conditions, the output priority is determined according to the risk value from high to low. When the number of drive channels is greater than or equal to the number of output areas, multiple areas are output in a time-sharing or partitioned manner. When the number of drive channels is less than the number of areas to be output, the area with the highest risk value is output first. During coil activation, coil temperature, drive current, operating time, and human body movement status are continuously monitored.

10. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, The closed-loop feedback adjustment includes: When the pressure risk value of the target area decreases after output, maintain the current parameters or gradually reduce the stimulation intensity; If the pressure risk value of the target area does not decrease after output and the safety limit is not triggered, adjust the stimulation parameters or expand the coverage of adjacent coils within the safe range. The magnetic stimulation output will automatically stop or pause when the system detects that a person is getting out of bed, turning over or sitting up, the coil temperature exceeds the safety threshold, the drive current is abnormal, the sensor data is abnormal, or the caregiver manually stops the stimulation.

11. The closed-loop control method for bone protuberance targeted magnetic stimulation based on a flexible pressure array as described in claim 1, characterized in that, It also includes the following steps: Nursing logs are generated by recording human posture, high-risk areas of bony prominences, pressure risk values, activation coil locations, magnetic stimulation parameters, output duration, safety status, and closed-loop adjustment results. The nursing terminal displays the current human lying position, pressure heat map, high-risk areas of bony prominences, the location of the working magnetic stimulation coil, and the risk change trend; When the risk value of the bony prominence continuously exceeds the preset risk threshold for more than the first preset time, the risk value does not decrease after intervention, the human body is not turned over for more than the second preset time, or the equipment is malfunctioning, a prompt message is sent to the nursing staff.

12. A closed-loop control system for bone protuberance-targeted magnetic stimulation based on a flexible pressure array, characterized in that, The system for performing the method as described in any one of claims 1 to 11 includes: A flexible pressure array sensing module is laid on the bed surface to continuously collect pressure distribution data of the human body while in bed. The data preprocessing module is used to perform zero-point calibration, outlier removal, smoothing filtering, and normalization on the collected pressure data to generate pressure characteristic data. The human body contact area extraction module is used to extract the effective contact area between the human body and the bed surface based on the pressure feature data, obtain the overall pressure profile of the human body on the bed surface, and generate a human body contact area map based on the effective contact area. The human posture reconstruction module is used to generate a pressure distribution map, a pressure gradient map, and a historical pressure change sequence based on the pressure feature data. The pressure distribution map, human contact area map, pressure gradient map, and historical pressure change sequence are input into a multi-task deep learning model to reconstruct the human contour posture and identify the lying posture type and candidate regions of bony landmarks. The bony prominence pressure area identification module is used to identify the bony prominence pressure area and calculate the bony prominence risk value to characterize the risk of pressure injury based on the posture reconstruction results and the current pressure distribution. The spatial mapping module is used to establish the correspondence between the pressure array coordinates and the magnetic stimulation coil coordinates, and to map the identified bony prominence pressure area to the corresponding magnetic induction coil unit or coil combination, generating control commands containing target coil position information; The magnetic stimulation parameter generation module is used to automatically generate magnetic stimulation parameters based on the type of bony prominence region and the risk value of the bony prominence. A flexible PCB magnetic induction coil array module is installed under the human body in a mattress. It includes multiple independently controllable magnetic induction coil units for outputting magnetic stimulation according to the control commands and magnetic stimulation parameters. The drive control module is used to drive the coil unit at the corresponding position to work, and to monitor the coil temperature, drive current and working time; The closed-loop feedback control module is used to dynamically adjust based on changes in pressure and bony prominence risk before and after stimulation, and to automatically stop or pause the magnetic stimulation output when an abnormal state is detected.

13. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 11.

Images