Sleep posture monitoring and intelligent pillow height adjusting method and system

Abstract

The invention provides a sleep posture monitoring and intelligent pillow height adjusting method and system, and relates to the technical field of intelligent sleep assisting devices.The method comprises the steps of obtaining head and neck contact pressure time sequence data, calculating a pressure distribution feature recognition posture mode, extracting cervical vertebra supporting area pressure features, and combining cervical vertebra physiological curvature parameters of a user; based on the constraint relation between the pressure distribution unevenness and the posture stability index, the optimal height adjusting combination of the pillow body supporting subareas is calculated, the multiple independent supporting units of the pillow body are driven to form a three-dimensional supporting curved surface matched with the head and neck posture, and therefore cervical vertebra stress is transferred to a bony supporting structure, and the sleep comfort is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent sleep aids, and in particular to a method and system for sleep posture monitoring and intelligent pillow height adjustment. Background Technology

[0002] Sleep quality is closely related to head and neck posture support, and an appropriate pillow height is crucial for maintaining the physiological curve of the cervical spine and reducing neck pressure. With increasing public awareness of sleep health, smart sleep aids have gradually become a research hotspot. Traditional pillows typically have a fixed height design, which cannot adapt to the physiological differences of different users and the needs of posture changes during sleep. In recent years, adjustable-height pillows and related monitoring technologies have made some progress, including pressure-sensor-based sleep posture recognition systems and mechanical structures that automatically adjust pillow height. Summary of the Invention

[0003] The present invention provides a method and system for sleep posture monitoring and intelligent pillow height adjustment, which can solve the problems in the prior art.

[0004] A first aspect of the present invention provides a method for sleep posture monitoring and intelligent pillow height adjustment, comprising: Acquire time-series data of head and neck contact pressure of the target user during the sleep cycle; Based on the time-series data of head and neck contact pressure, the pressure distribution matrix is ​​calculated to obtain the position of the pressure centroid and its migration trajectory. Combined with the morphological change characteristics of the pressure peak region, the head and neck posture transition mode is identified. At the same time, the posture stability index is calculated based on the spatial gradient and temporal evolution characteristics of the pressure distribution. Based on the posture transition pattern, the sleep cycle is divided into a posture transition phase and a posture stabilization phase. For the posture stabilization phase, the pressure topology skeleton of the cervical spine support area is extracted, and its topological complexity, density characteristics and pressure propagation steepness are calculated. The pressure distribution unevenness index is calculated in combination with the pressure range. During the initial usage period, the personalized cervical spine physiological curvature parameters of the target user are extracted; Based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, and combined with the cervical spine physiological curvature parameters, the pressure distribution direction is optimized by adjusting the height gradient between adjacent support zones of the pillow body, so that the force on the cervical spine is transferred to the bony support structure, and the optimal height adjustment combination of different support zones of the pillow body is calculated. The height adjustment combination drives multiple independent support units within the pillow to perform differentiated height changes, forming a three-dimensional support surface that adapts to the current head and neck posture.

[0005] Based on the time-series data of head and neck contact pressure, a pressure distribution matrix is ​​calculated to obtain the location of the pressure centroid and its migration trajectory. Combined with the morphological changes in the pressure peak region, head and neck posture transition patterns are identified. Simultaneously, based on the spatial gradient and temporal evolution characteristics of the pressure distribution, posture stability indices are calculated, including: The time-series data of head and neck contact pressure is sampled and segmented at fixed time intervals. The position of the pressure centroid of the pressure distribution matrix in each time period is calculated, and the migration trajectory is obtained by connecting them in time sequence. The pressure values ​​in the pressure distribution matrix are converted into pressure potential energy per unit area. The pressure potential energy gradient between adjacent pressure sampling points is calculated. An energy flow vector field characterizing the pressure transmission path and intensity distribution is generated. The energy inflow-outflow flux ratio in the pressure peak region is calculated to obtain the energy accumulation rate. Extract the equivalent radius of the pressure peak region, calculate the correlation coefficient between the equivalent radius and the energy accumulation rate, and determine that a head and neck attitude transition has occurred when the correlation coefficient is negative and its absolute value exceeds the attitude determination threshold. Based on the migration trajectory and the principal direction of the energy flow vector field, determine the head and neck attitude transition mode. The gradient distribution of the pressure distribution matrix in the horizontal, vertical and diagonal directions is calculated, the standard deviation and extreme point density of the gradient in each direction are extracted as spatial feature parameters, the local change frequency of the pressure distribution matrix is ​​calculated, and the pressure time series data is decomposed into fast-changing components and slow-changing components. The spatial characteristic parameters are multiplied by the energy entropy of the rapidly changing component, and then differentially fused with the trend change of the slowly changing component to generate the attitude stability index.

[0006] Based on the posture transition pattern, the sleep cycle is divided into a posture transition phase and a posture stabilization phase. For the posture stabilization phase, the pressure topology skeleton of the cervical spine support area is extracted, and its topological complexity, density characteristics, and pressure propagation steepness are calculated. Combined with the pressure range, the pressure distribution unevenness index is calculated, including: The time of attitude transition in the attitude transition mode is marked as the transition boundary point of the sleep cycle. The time integral value of the attitude stability index between adjacent transition boundary points is calculated. The time period when the integral value is higher than the stability integral threshold is marked as the attitude stability stage, and the rest is marked as the attitude transition stage. For the posture stabilization phase, the cervical spine support area is extracted and its pressure distribution is morphologically refined. A pressure topology skeleton with a single pixel width is extracted, the number of branch points is counted to obtain the topology complexity, and the ratio of the total skeleton length to the region area is calculated to obtain the density feature. The pressure topology skeleton is traced from the high pressure point to the low pressure point, the pressure drop rate is calculated, and the maximum pressure drop rate is used as the stress propagation steepness. The cervical spine support area is divided into grid cells, the pressure range within each grid cell is calculated and weighted summation is performed, and the pressure distribution non-uniformity index is obtained by combining the topological complexity, the density characteristics and the stress propagation steepness.

[0007] For the posture stabilization phase, the cervical spine support region is extracted and its pressure distribution is morphologically refined. A single-pixel width pressure topology skeleton is extracted, the number of branch points is counted to obtain the topological complexity, and the density feature is obtained by calculating the ratio of the total skeleton length to the region area, including: Based on the pressure distribution data during the posture stabilization phase, a binary pressure image of the cervical spine contact area is segmented by setting a pressure threshold, which serves as the initial boundary of the cervical spine support area. Morphological thinning is performed on the binarized stress image. Under the constraints of maintaining regional connectivity and topology, the outer pixels are peeled off layer by layer. For each pixel to be deleted, connectivity is checked and endpoint protection is determined to avoid regional splitting or the generation of isolated pixels, until a continuous and complete single-pixel width stress topology skeleton is obtained. Traverse all pixels of the pressure topology skeleton, identify pixels with more than the critical number of adjacent pixels as branch points, count the total number of branch points to obtain the topology complexity, calculate the cumulative distance between adjacent pixels based on the skeleton pixel coordinate sequence to obtain the total skeleton length, count the number of all effective pixels in the cervical spine support area as the area, and calculate the ratio of the total skeleton length to the area to obtain the skeleton distribution density per unit area, which is used as the density feature.

[0008] During the initial usage period, the personalized cervical spine physiological curvature parameters of the target user are extracted, including: During multiple posture stabilization phases in the initial use period, pressure distribution data of the cervical spine support area are collected, spatial interpolation is performed on the pressure distribution data to generate a continuous pressure field, isobar clusters of the continuous pressure field are calculated, and the closed contours and centroids of each isobar are extracted. The centroids of isobars under different pressure values ​​are connected from high to low pressure to form a centroid series curve. The point of maximum curvature is identified as the projection position of the cervical spine curvature vertex. The curve is divided into the occipital bone support segment and the cervical spine support segment. The average radius of curvature and segment length ratio of the two segments are calculated. The distribution of the average radius of curvature and the ratio of segment length within different sleep days was statistically analyzed. The median of each parameter was calculated, outliers with interquartile ranges exceeding the median tolerance were removed, and the remaining data were weighted and averaged to obtain the stable radius of curvature and the stable ratio of segment length. The standard curvature distribution of the cervical spine support segment is determined based on the stable curvature radius, and its range of action in the occipital support area is determined by the proportion of the stable segment length. The peak position of the curvature distribution is determined by the projection position of the cervical spine flexure vertex, thus obtaining the personalized cervical spine physiological curvature parameters of the target user.

[0009] Based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, and combined with the cervical spine physiological curvature parameters, the pressure distribution direction is optimized by adjusting the height gradient between adjacent support zones of the pillow body, so that the force on the cervical spine is transferred to the bony support structure. The optimal height adjustment combination for different support zones of the pillow body is calculated, including: Using the pressure distribution non-uniformity index as a spatial constraint variable and the posture stability index as a time constraint variable, a bivariate correlation function characterizing the coupling constraint relationship is constructed; the curvature information of each segment in the cervical spine physiological curvature parameters is extracted and mapped to each support zone of the occipital body to determine the location of the cervical spine bony support structure and the soft tissue distribution area; Based on the cervical spine physiological curvature parameters, the target height difference between adjacent support zones is calculated to form a height gradient distribution. A local coordinate system is established to calculate the normal pressure component and tangential pressure component at each point on the pillow surface to determine the pressure distribution direction. Increase the height of the corresponding zone of the cervical vertebral bony support structure and decrease the height of the corresponding zone of the soft tissue distribution area to form a gradient field that guides the pressure distribution direction to be transmitted to the bony support structure; Calculate the distribution ratio of cervical spine force in the bony support structure under the action of the gradient field. When the distribution ratio reaches the preset force transmission threshold, record the height value of each support zone. Substitute the height value into the bivariate correlation function for verification. When the constraint conditions are met, determine it as the optimal height adjustment combination.

[0010] The height adjustment combination drives multiple independent support units within the pillow body to perform differentiated height changes, forming a three-dimensional support surface adapted to the current head and neck posture, including: Based on the height adjustment combination, the adjustment difference between the target height and the current height of each support zone is calculated. When the height adjustment difference between adjacent support zones exceeds the preset sudden change limit, a transition height sequence is calculated based on linear interpolation. The transition sequence is then allocated to the independent support units between adjacent support zones to achieve continuous height transition. Based on the attitude stability index, the adjustment rate of each independent support unit is calculated, and the target height and adjustment rate are sent to each independent support unit. Pressure timing data is collected during the lifting and lowering process. The pressure change caused by the unit height change is calculated to obtain the pressure response coefficient. When the pressure response coefficient exceeds the pressure change threshold, the system returns to the previous height position and recalculates the adjustment rate. The process continues until the target height is reached. After each independent support unit reaches the target height, a three-dimensional support surface that adapts to the head and neck posture is formed.

[0011] A second aspect of the present invention provides a sleep posture monitoring and intelligent pillow height adjustment system, comprising: The data acquisition unit is used to acquire time-series data of head and neck contact pressure of the target user during the sleep cycle; The posture analysis and stability assessment unit is used to calculate the pressure distribution matrix based on the time series data of head and neck contact pressure, obtain the position of the pressure centroid and its migration trajectory, identify the head and neck posture transition mode by combining the morphological change characteristics of the pressure peak region, and calculate the posture stability index according to the spatial gradient and temporal evolution characteristics of the pressure distribution. The pressure distribution non-uniformity calculation unit is used to divide the sleep cycle into a posture transition stage and a posture stability stage according to the posture transition mode. For the posture stability stage, the pressure topology skeleton of the cervical spine support area is extracted, and its topology complexity, density characteristics and pressure propagation steepness are calculated. The pressure distribution non-uniformity index is calculated in combination with the pressure range. The cervical curvature parameter extraction unit is used to extract the personalized cervical physiological curvature parameters of the target user within the initial usage period. The support height optimization calculation unit is used to optimize the pressure distribution direction by adjusting the height gradient between adjacent support zones of the pillow body based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, combined with the cervical spine physiological curvature parameters, so that the force on the cervical spine is transferred to the bony support structure, and calculate the optimal height adjustment combination of different support zones of the pillow body. The dynamic adjustment execution unit is used to drive multiple independent support units within the pillow body to perform differentiated height changes according to the height adjustment combination, forming a three-dimensional support surface that adapts to the current head and neck posture.

[0012] A third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0013] Fourth aspect of the present invention, A computer-readable storage medium is provided, having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0014] The beneficial effects of this application are as follows: By acquiring time-series data of head and neck contact pressure and calculating the pressure distribution matrix and the migration trajectory of the pressure centroid, accurate identification of head and neck posture transition patterns is achieved, which can automatically adapt to the sleep habits of different users.

[0015] Based on the spatial gradient and temporal evolution characteristics of pressure distribution, a posture stability index is calculated, which can scientifically assess the stability of a user's sleep posture and provide a data foundation for intelligent adjustment. 3. By dividing the posture into a transitional phase and a stable phase, the pressure topology skeleton of the cervical spine support area is extracted for the stable phase, and relevant feature parameters are calculated, achieving a more accurate assessment of head and neck support needs.

[0016] By combining personalized cervical spine physiological curvature parameters with the coupling constraint relationship between pressure distribution unevenness index and postural stability index, an adjustment model is constructed, avoiding the limitations of traditional fixed pillow height schemes. This method achieves dynamic monitoring and intelligent adjustment during sleep, and compared with traditional fixed pillow height or manual adjustment schemes, it has higher human-machine coordination and adaptability, significantly improving neck discomfort caused by sleep posture. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the sleep posture monitoring and intelligent pillow height adjustment method according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating the optimal height adjustment combination generation method in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0020] Figure 1 This is a flowchart illustrating the sleep posture monitoring and intelligent pillow height adjustment method according to an embodiment of the present invention. Figure 1As shown, the method includes: Acquire time-series data of head and neck contact pressure of the target user during the sleep cycle; Based on the time-series data of head and neck contact pressure, the pressure distribution matrix is ​​calculated to obtain the position of the pressure centroid and its migration trajectory. Combined with the morphological change characteristics of the pressure peak region, the head and neck posture transition mode is identified. At the same time, the posture stability index is calculated based on the spatial gradient and temporal evolution characteristics of the pressure distribution. Based on the posture transition pattern, the sleep cycle is divided into a posture transition phase and a posture stabilization phase. For the posture stabilization phase, the pressure topology skeleton of the cervical spine support area is extracted, and its topological complexity, density characteristics and pressure propagation steepness are calculated. The pressure distribution unevenness index is calculated in combination with the pressure range. During the initial usage period, the personalized cervical spine physiological curvature parameters of the target user are extracted; Based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, and combined with the cervical spine physiological curvature parameters, the pressure distribution direction is optimized by adjusting the height gradient between adjacent support zones of the pillow body, so that the force on the cervical spine is transferred to the bony support structure, and the optimal height adjustment combination of different support zones of the pillow body is calculated. The height adjustment combination drives multiple independent support units within the pillow to perform differentiated height changes, forming a three-dimensional support surface that adapts to the current head and neck posture.

[0021] In one optional implementation, a pressure distribution matrix is ​​calculated based on the time-series data of head and neck contact pressure to obtain the location of the pressure centroid and its migration trajectory. Combined with the morphological change characteristics of the pressure peak region, head and neck posture transition patterns are identified. Simultaneously, posture stability indices are calculated based on the spatial gradient and temporal evolution characteristics of the pressure distribution, including: The time-series data of head and neck contact pressure is sampled and segmented at fixed time intervals. The position of the pressure centroid of the pressure distribution matrix in each time period is calculated, and the migration trajectory is obtained by connecting them in time sequence. The pressure values ​​in the pressure distribution matrix are converted into pressure potential energy per unit area. The pressure potential energy gradient between adjacent pressure sampling points is calculated. An energy flow vector field characterizing the pressure transmission path and intensity distribution is generated. The energy inflow-outflow flux ratio in the pressure peak region is calculated to obtain the energy accumulation rate. Extract the equivalent radius of the pressure peak region, calculate the correlation coefficient between the equivalent radius and the energy accumulation rate, and determine that a head and neck attitude transition has occurred when the correlation coefficient is negative and its absolute value exceeds the attitude determination threshold. Based on the migration trajectory and the principal direction of the energy flow vector field, determine the head and neck attitude transition mode. The gradient distribution of the pressure distribution matrix in the horizontal, vertical and diagonal directions is calculated, the standard deviation and extreme point density of the gradient in each direction are extracted as spatial feature parameters, the local change frequency of the pressure distribution matrix is ​​calculated, and the pressure time series data is decomposed into fast-changing components and slow-changing components. The spatial characteristic parameters are multiplied by the energy entropy of the rapidly changing component, and then differentially fused with the trend change of the slowly changing component to generate the attitude stability index.

[0022] In this specific embodiment, it is necessary to collect time-series data of head and neck contact pressure. This data is typically acquired through a pressure sensor array placed on a supporting surface such as a pillow, mattress, or neck brace, recording the pressure distribution data generated when the human head and neck come into contact with the supporting surface.

[0023] The collected pressure time-series data is sampled in segments at fixed time intervals, for example, every 0.5 seconds. For each time segment, an m×n dimensional pressure distribution matrix P is constructed, where the matrix element P(i,j) represents the pressure value at position (i,j). Based on this matrix, the position of the pressure centroid (C) is calculated. x C y ): The horizontal axis of the pressure centroid C x The pressure is equal to the sum of the x-coordinates of each point multiplied by the corresponding pressure values, divided by the total pressure; the y-coordinate of the pressure centroid is C. y The sum of the ordinates of each point multiplied by the corresponding pressure values ​​divided by the total pressure is equal to the total pressure. Connecting the pressure centroid positions at different time points in chronological order forms the centroid migration trajectory T, which reflects the change in the head and neck position over time.

[0024] The pressure values ​​in the pressure distribution matrix are converted into pressure potential energy per unit area E(i,j). The conversion method is to square the pressure value P(i,j) and multiply it by a scaling factor k (usually taken as 0.5), that is, E(i,j)=k×P(i,j) 2 Calculate the pressure potential energy gradient G(i,j) between adjacent pressure sampling points, which represents the rate of change of potential energy in the horizontal and vertical directions. Based on the potential energy gradient, generate an energy flow vector field V, which points in the direction of decreasing potential energy and whose magnitude is proportional to the gradient.

[0025] Determine the pressure peak region R, which is a continuous region where the pressure value exceeds a set threshold (such as 70% of the maximum pressure value). Calculate the energy inflow and outflow fluxes within region R. The inflow flux Fin is the total energy flow pointing into the region, and the outflow flux Fout is the total energy flow pointing out of the region. Calculate the energy accumulation rate AR = Fin / Fout, which reflects the energy accumulation state in the pressure peak region.

[0026] The equivalent radius ER of the pressure peak region R is extracted by converting the region area into the radius of a circle with equal area. The Pearson correlation coefficient CR between the equivalent radius ER and the energy accumulation rate AR is calculated. When CR is negative and its absolute value exceeds a preset attitude determination threshold (e.g., 0.6), a head and neck attitude transition is determined to have occurred. A negative correlation indicates that as the pressure region expands, the energy accumulation rate decreases, signifying a pressure redistribution process.

[0027] Based on the centroid migration trajectory T and the principal direction of the energy flow vector field V (obtained through principal component analysis of the vector field), the head and neck posture transition mode is determined. For example, when the centroid moves to the right and the principal direction of energy flow is to the right, it is determined to be turning the head to the right; when the centroid moves upward and the principal direction of energy flow is upward, it is determined to be a head-raising action.

[0028] Calculate the gradient distribution G of the pressure distribution matrix in the horizontal, vertical, and diagonal directions. x G y G d Extract the standard deviation (SD) of the gradient in each direction. x SD y SD d As a characteristic of gradient distribution, the density of extreme points DP of gradients in each direction is calculated. x DP y DP d , representing the number of gradient extrema points per unit area, these parameters together constitute the spatial feature parameter set SF={SD x SD y SD d DP x DP y DP d}

[0029] The local variation frequency LF of the pressure distribution matrix is ​​calculated by analyzing the distribution of the rate of change of pressure values ​​at the same location within adjacent time periods. Wavelet decomposition or empirical mode decomposition methods are used to decompose the pressure time series data into a fast-varying component HF (high-frequency part) and a slow-varying component LF (low-frequency part). The fast-varying component reflects the pressure fluctuations in a short period of time, while the slow-varying component reflects the long-term trend.

[0030] The energy entropy EE of the rapidly varying component HF is calculated to characterize the complexity of pressure fluctuations. The spatial characteristic parameters are multiplied by the energy entropy of the rapidly varying component to obtain the parameter SF×EE. The trend change TC of the slowly varying component LF is calculated, representing the slope of the linear fit of the slowly varying component within a continuous time window.

[0031] The SF×EE and TC are differentially fused to generate the attitude stability index SI = α×(SF×EE)-β×TC, where α and β are weighting coefficients, typically α=0.7 and β=0.3. The lower the SI value, the more stable the head and neck posture; the higher the SI value, the stronger the attitude instability.

[0032] In practical applications, this method can be used to monitor changes in head and neck posture in bedridden patients, assess sleep quality, or assist in posture correction for patients with cervical spondylosis. For example, by continuously monitoring the head and neck contact pressure distribution when patients with cervical spondylosis use a neck brace, it can identify poor posture and provide timely feedback, helping patients maintain correct head and neck posture and alleviate symptoms.

[0033] This method can also be applied to smart pillow systems, automatically adjusting the pillow's shape or firmness based on pressure distribution and postural stability indicators to provide users with more ergonomic support and improve sleep comfort and quality. Through long-term data collection and analysis, personalized head and neck posture assessment models can also be established, providing a scientific basis for rehabilitation training.

[0034] In one optional implementation, based on the posture transition pattern, the sleep cycle is divided into a posture transition phase and a posture stabilization phase. For the posture stabilization phase, the pressure topology skeleton of the cervical spine support area is extracted, and its topological complexity, density characteristics, and pressure propagation steepness are calculated. The pressure distribution unevenness index is calculated in conjunction with the pressure range, including: The time of attitude transition in the attitude transition mode is marked as the transition boundary point of the sleep cycle. The time integral value of the attitude stability index between adjacent transition boundary points is calculated. The time period when the integral value is higher than the stability integral threshold is marked as the attitude stability stage, and the rest is marked as the attitude transition stage. For the posture stabilization phase, the cervical spine support area is extracted and its pressure distribution is morphologically refined. A pressure topology skeleton with a single pixel width is extracted, the number of branch points is counted to obtain the topology complexity, and the ratio of the total skeleton length to the region area is calculated to obtain the density feature. The pressure topology skeleton is traced from the high pressure point to the low pressure point, the pressure drop rate is calculated, and the maximum pressure drop rate is used as the stress propagation steepness. The cervical spine support area is divided into grid cells, the pressure range within each grid cell is calculated and weighted summation is performed, and the pressure distribution non-uniformity index is obtained by combining the topological complexity, the density characteristics and the stress propagation steepness.

[0035] In this specific embodiment, based on the posture transition pattern, the sleep cycle is divided into a posture transition phase and a posture stabilization phase. By identifying the occurrence time of posture transitions, these times are marked as transition boundary points of the sleep cycle. Between two adjacent transition boundary points, the time integral value of the posture stability index is calculated. The posture stability index can be obtained by comprehensively considering parameters such as the displacement velocity of the pressure center of gravity and the rate of change of pressure distribution. For example, when the user's posture remains relatively stable, the displacement velocity of the pressure center of gravity is low, and the rate of change of pressure distribution is small. The posture stability index is integrated over time and compared with a preset stability integral threshold. When the integral value is higher than the stability integral threshold, the time period is marked as the posture stabilization phase; otherwise, it is marked as the posture transition phase.

[0036] For periods marked as posturally stable, pressure distribution data were further analyzed. Based on ergonomic principles and anatomical features, combined with the current sleep posture, the cervical spine support area was located. In the supine position, the cervical spine support area is typically located in the area where the pillow contacts the neck; in the lateral position, this area is located in the area where the pillow contacts the side of the head and the side of the neck.

[0037] The extracted pressure distribution data of the cervical spine support region is subjected to morphological refinement to retain the main structural features of the pressure distribution and remove noise and redundant information. Through the morphological refinement algorithm, the pressure distribution of the cervical spine support region is transformed into a pressure topology skeleton with a width of one pixel, which retains the topological structural features of the pressure distribution.

[0038] After obtaining the pressure topology skeleton, the number of branch points in the skeleton is counted as a quantitative indicator of topological complexity. A branch point is a position in the skeleton where one pixel connects to three or more other pixels, representing the complexity of the pressure distribution. Higher topological complexity indicates stronger irregularity in the pressure distribution. Simultaneously, the ratio of the total skeleton length to the area of ​​the cervical spine support region is calculated to obtain the density feature. The density feature reflects the complexity of the pressure distribution per unit area; a higher density value indicates a more complex pressure distribution.

[0039] Tracing the pressure topology from high-pressure point to low-pressure point, the pressure changes along the path are recorded. High-pressure points are typically the points with the highest pressure values ​​on the skeleton, while low-pressure points are the points with the lowest pressure values ​​among the skeleton's endpoints. The pressure drop rate between adjacent points along the path is calculated, and the maximum pressure drop rate is identified as the stress propagation steepness. The stress propagation steepness reflects the pressure transmission characteristics in the cervical spine support region. A greater steepness indicates a more uneven pressure distribution, which can lead to excessively high local pressure.

[0040] The cervical spine support area is divided into several grid cells, such as 5×5 or 8×8 grids. For each grid cell, the pressure range within it is calculated, which is the difference between the maximum and minimum pressure values. Based on the importance of each grid cell in cervical spine support, a weight coefficient is assigned to each cell. For example, the grid cell containing the direct support point of the cervical curve can be assigned a higher weight. The pressure range of each grid cell is multiplied by its corresponding weight and summed to obtain the weighted pressure range.

[0041] Combining the previously calculated topological complexity, density characteristics, and stress propagation steepness, a pressure distribution non-uniformity index is calculated using a weighted fusion method. A linear weighting method can be adopted: Non-uniformity index = α × topological complexity + β × density characteristics + γ × stress propagation steepness + δ × weighted pressure range, where α, β, γ, and δ are weighting coefficients that can be adjusted according to the actual application scenario and requirements.

[0042] In practical applications, the pressure distribution unevenness index can be used to assess the health of sleeping posture and the effectiveness of pillow support. For example, when the index value exceeds a preset threshold, it can prompt the user to adjust their sleeping posture or change to a more suitable pillow to improve cervical spine support and reduce sleep-related neck discomfort.

[0043] By conducting a detailed analysis of the pressure distribution in the cervical spine support area during the postural stabilization phase, the stress on the cervical spine during sleep can be quantitatively assessed, providing an objective basis for sleep health management. Furthermore, continuous monitoring of the changing trends of pressure distribution unevenness indicators can also help assess the long-term impact of using specific pillows on cervical spine health, providing data support for personalized sleep product recommendations.

[0044] In one optional implementation, for the posture stabilization phase, the cervical spine support region is extracted and its pressure distribution is morphologically refined. A pressure topology skeleton with a single pixel width is extracted, the number of branch points is counted to obtain the topological complexity, and the density feature is obtained by calculating the ratio of the total skeleton length to the region area. This includes: Based on the pressure distribution data during the posture stabilization phase, a binary pressure image of the cervical spine contact area is segmented by setting a pressure threshold, which serves as the initial boundary of the cervical spine support area. Morphological thinning is performed on the binarized stress image. Under the constraints of maintaining regional connectivity and topology, the outer pixels are peeled off layer by layer. For each pixel to be deleted, connectivity is checked and endpoint protection is determined to avoid regional splitting or the generation of isolated pixels, until a continuous and complete single-pixel width stress topology skeleton is obtained. Traverse all pixels of the pressure topology skeleton, identify pixels with more than the critical number of adjacent pixels as branch points, count the total number of branch points to obtain the topology complexity, calculate the cumulative distance between adjacent pixels based on the skeleton pixel coordinate sequence to obtain the total skeleton length, count the number of all effective pixels in the cervical spine support area as the area, and calculate the ratio of the total skeleton length to the area to obtain the skeleton distribution density per unit area, which is used as the density feature.

[0045] In this specific embodiment, pressure distribution data during the posture stabilization phase is acquired. The data is collected by a pressure sensing array while the subject maintains a stable posture. Each sampling point records the pressure value at the corresponding location. Based on this data, an appropriate pressure threshold (such as 20% of the maximum pressure value) is set to segment the cervical spine contact area. When the pressure value of a sampling point is greater than or equal to the set threshold, it is marked as 1 (foreground) in the binary pressure image; otherwise, it is marked as 0 (background). In this way, an initial binary image representing the cervical spine support area can be obtained.

[0046] Morphological thinning is performed on the binarized pressure image. The thinning process is based on the principle of iterative boundary pixel stripping. Boundary pixels are removed layer by layer under the constraint of maintaining regional connectivity and topological characteristics. The specific thinning algorithm adopts the following steps: Define the eight-neighborhood connection relationship, that is, each pixel forms a connection relationship with its adjacent pixels in the above, below, left, right and four diagonal directions. In each iteration, the boundary pixels that meet the following conditions are marked: (1) The pixel is a foreground pixel; (2) At least one eight-neighborhood pixel is the background; (3) Removing the pixel will not cause the region to split; (4) The pixel is not an endpoint (an endpoint is defined as a pixel with only one foreground pixel in the eight-neighborhood).

[0047] In the connectivity test, the local connectivity number calculation method is used to determine whether a pixel is a connected point. For a pixel P, the number of transitions from 0 to 1 in the binary sequence of its eight neighbors arranged clockwise or counterclockwise is checked. If the number of transitions is not 1, removing the pixel will change the topology of the region, so it should be retained. At the same time, in order to avoid skeleton breakage, a protection strategy is implemented for the endpoints, that is, points with only one foreground pixel in their eight neighbors are not deleted.

[0048] The above process is repeated iteratively until no pixels can be further deleted, resulting in a single-pixel-wide pressure topology skeleton. This skeleton retains the basic morphological features and topological structure of the original region, but greatly simplifies the data representation.

[0049] After the skeleton is extracted, skeleton feature analysis is performed. By traversing each pixel on the skeleton, its connection characteristics are identified. The critical adjacency number is defined as 2. That is, when the number of foreground pixels in the eight neighbors of a pixel is greater than 2, it is determined as a branch point. The total number of all branch points in the skeleton is counted to obtain the topological complexity index that characterizes the complexity of the cervical spine support region.

[0050] In addition, the total skeleton length is calculated as a morphological feature by accumulating the Euclidean distance between adjacent pixels along the skeleton path: for horizontally or vertically adjacent pixel pairs, the distance is recorded as 1 unit; for diagonally adjacent pixel pairs, the distance is recorded as 1.414 units. The total skeleton length is obtained by accumulating these local distances along the skeleton path.

[0051] At the same time, the number of all foreground pixels in the cervical spine support area is counted as the area index. The total skeleton length calculated above is divided by the area to obtain the skeleton distribution density per unit area. This density feature reflects the concentration and complexity of pressure distribution in the cervical spine support area.

[0052] In practical applications, when processing pressure data from different subjects or different postures, the pressure threshold can be dynamically adjusted according to actual needs. For example, for subjects with lighter weight, the threshold can be appropriately lowered to ensure complete capture of the cervical spine support area. For specific posture analysis, the optimal segmentation threshold can be determined by observing the pressure distribution histogram. The topological complexity and density features obtained by the above methods can effectively quantify the pressure distribution characteristics of the cervical spine support area.

[0053] In one optional implementation, during the initial usage period, the personalized cervical curvature parameters of the target user are extracted, including: During multiple posture stabilization phases in the initial use period, pressure distribution data of the cervical spine support area are collected, spatial interpolation is performed on the pressure distribution data to generate a continuous pressure field, isobar clusters of the continuous pressure field are calculated, and the closed contours and centroids of each isobar are extracted. The centroids of isobars under different pressure values ​​are connected from high to low pressure to form a centroid series curve. The point of maximum curvature is identified as the projection position of the cervical spine curvature vertex. The curve is divided into the occipital bone support segment and the cervical spine support segment. The average radius of curvature and segment length ratio of the two segments are calculated. The distribution of the average radius of curvature and the ratio of segment length within different sleep days was statistically analyzed. The median of each parameter was calculated, outliers with interquartile ranges exceeding the median tolerance were removed, and the remaining data were weighted and averaged to obtain the stable radius of curvature and the stable ratio of segment length. The standard curvature distribution of the cervical spine support segment is determined based on the stable curvature radius, and its range of action in the occipital support area is determined by the proportion of the stable segment length. The peak position of the curvature distribution is determined by the projection position of the cervical spine flexure vertex, thus obtaining the personalized cervical spine physiological curvature parameters of the target user.

[0054] In this specific embodiment, during the initial use phase, pressure distribution data of the cervical spine support area during multiple stable sleep postures are collected. Specifically, a pressure sensor array is arranged inside the pillow, and each sensor unit records the pressure value at a specific location, forming a discrete pressure distribution matrix. When the user enters a stable sleep state, that is, when the head position remains relatively fixed and breathing is stable, the data acquisition process is triggered, and pressure data for at least five minutes is recorded and the average value is calculated to obtain the pressure distribution data under the stable posture.

[0055] Spatial interpolation is performed on the collected discrete pressure distribution data to generate a continuous pressure field. The radial basis function interpolation method is used to extend the discrete sensor data points into a continuous distribution function, ensuring that any location within the interpolation region has a corresponding pressure value. The formula for generating the continuous pressure field can be expressed as: For any point (x, y) within any region, its pressure value can be obtained by using the known sensor location (x, y). i y i The pressure value Pi is calculated using a weighting function.

[0056] Based on the generated continuous pressure field, isobars are calculated. Starting from the maximum pressure value, multiple isobars are generated in an arithmetic descending manner. The common isobar difference is set to 8%-10% of the maximum pressure value, and usually 1015 isobars are generated. The isobars are calculated using a contour tracing algorithm. After discretizing the continuous pressure field into a grid, the closed contour lines under each isobar are determined in sequence.

[0057] For each closed isobar, its centroid position is extracted. The centroid calculation of a closed isobar is based on the geometry of the contour line. For contours with relatively regular shapes, the geometric center can be directly used as the centroid. For contours with irregular shapes, the centroid position is calculated by weighting the pressure within the contour area.

[0058] Connect the centroids of isobars under different pressure values ​​from high to low pressure to form a centroid series curve. This curve reflects the pressure transition relationship from the cervical spine support point to the occipital support area. Its shape is closely related to the curvature of the user's cervical spine. Calculate the curvature distribution of the generated centroid series curve and identify the point with the most significant curvature change, i.e., the maximum curvature point. This point usually corresponds to the projection position of the cervical spine curvature vertex on the support surface.

[0059] Using the point of maximum curvature as the boundary, the centroid-connected curve is divided into two parts: the occipital support segment and the cervical support segment. The average radius of curvature and length of each segment are calculated, along with the segment length ratio. The radius of curvature is calculated using the three-point method: three adjacent points are selected on the curve, and a circle is defined by these three points; the radius of this circle is the radius of curvature at each point. The calculation window is moved along the curve to obtain the radius of curvature at each point, and the average value is taken as the average radius of curvature for that segment. The segment length ratio is expressed as the ratio of the length of the cervical support segment to the overall length of the centroid-connected curve.

[0060] During continuous use by the user, the distribution of the average radius of curvature and the ratio of segment length is statistically analyzed daily. For the collected multi-day data, the median of the ratio of radius of curvature and segment length is calculated, and a reasonable tolerance range is set. Usually, 1.5 times the interquartile range is taken as the tolerance multiple. Data points that exceed the median ± tolerance range are marked as outliers and removed.

[0061] The remaining valid data are processed by weighted average, where the weight coefficient is determined based on the sleep stability score corresponding to the data point. The sleep stability score mainly considers factors such as the duration of the stable phase, the coefficient of variation of the pressure distribution, and the frequency of the user turning over. The stable radius of curvature and the proportion of stable segment length are calculated by weighted average. These two parameters reflect the stable cervical spine support characteristics formed by the user during long-term use.

[0062] The standard curvature distribution of the cervical spine support segment is determined based on the stable curvature radius. The standard curvature distribution is usually fitted with a Gaussian function, and its peak position corresponds to the projection position of the cervical spine curvature vertex determined above. Combined with the proportion of the stable segment length, the range of effect of the curvature distribution in the occipital support area is determined, thereby fully describing the user's personalized cervical spine physiological curvature parameters.

[0063] In practical applications, the above parameters can be converted into pillow shape control points. For example, a user might extract the following cervical spine physiological curvature parameters: cervical support segment curvature radius 85mm, segment length ratio 1.2, and the apex of the cervical curve located 15mm posterior to the center of the support area. A smart pillow designed based on these parameters provides maximum support height at this location and gradually transitions to the occipital bone support area according to the calculated curvature distribution, thereby precisely matching the user's cervical spine physiological curve, effectively reducing neck pressure, and improving sleep quality.

[0064] For users who sleep on their side, the pressure distribution in the cervical spine support area exhibits different characteristics, typically showing more concentrated pressure peaks and steeper pressure gradients. In this case, the side-lying posture can be determined by identifying the lateral offset characteristics of the pressure center point, and the calculation method of the curvature parameters can be adjusted accordingly to ensure the accuracy of the parameters.

[0065] Figure 2This is a flowchart illustrating the method for generating the optimal height adjustment combination in an embodiment of the present invention. In an optional embodiment, based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, and combined with the cervical spine physiological curvature parameters, the pressure distribution direction is optimized by adjusting the height gradient between adjacent support zones of the pillow body, so that the force on the cervical spine is transferred to the bony support structure. The optimal height adjustment combination for different support zones of the pillow body is calculated, including: Using the pressure distribution non-uniformity index as a spatial constraint variable and the posture stability index as a time constraint variable, a bivariate correlation function characterizing the coupling constraint relationship is constructed; the curvature information of each segment in the cervical spine physiological curvature parameters is extracted and mapped to each support zone of the occipital body to determine the location of the cervical spine bony support structure and the soft tissue distribution area; Based on the cervical spine physiological curvature parameters, the target height difference between adjacent support zones is calculated to form a height gradient distribution. A local coordinate system is established to calculate the normal pressure component and tangential pressure component at each point on the pillow surface to determine the pressure distribution direction. Increase the height of the corresponding zone of the cervical vertebral bony support structure and decrease the height of the corresponding zone of the soft tissue distribution area to form a gradient field that guides the pressure distribution direction to be transmitted to the bony support structure; Calculate the distribution ratio of cervical spine force in the bony support structure under the action of the gradient field. When the distribution ratio reaches the preset force transmission threshold, record the height value of each support zone. Substitute the height value into the bivariate correlation function for verification. When the constraint conditions are met, determine it as the optimal height adjustment combination.

[0066] In this specific embodiment, the pressure distribution non-uniformity index is used as a spatial constraint variable, and the attitude stability index is used as a time constraint variable to construct a bivariate correlation function that characterizes the coupling constraint relationship. The pressure distribution non-uniformity index can be calculated from the data collected by the pressure sensor array. The bivariate correlation function can be expressed as F(P, S), where P represents the pressure distribution non-uniformity and S represents the attitude stability index. When F(P, S) ≤ K (K is a preset threshold), the coupling constraint condition is considered to be satisfied.

[0067] Curvature information of each segment in the cervical spine physiological curvature parameters is extracted and mapped to the occipital body support zones to determine the location of the bony support structures and the distribution area of ​​soft tissues. By analyzing standard cervical spine physiological curvature models or personalized cervical spine curve data, curvature information for each segment from C1 to C7 can be obtained. Projecting the cervical spine onto the occipital body surface, the occipital body can be divided into multiple functional support zones, typically including four main zones: the occipital region, upper cervical region, middle cervical region, and lower cervical region. Based on this, according to the anatomical location of the cervical spine skeletal structures, the bony support structures are mainly located in the occipital tuberosity, the posterior arch of the C1 vertebral body, and the occipital region and lower cervical region corresponding to the spinous process of C7, while the cervical soft tissues are mainly distributed in the upper and middle cervical regions.

[0068] Based on the cervical spine's physiological curvature parameters, the target height difference between adjacent support zones is calculated to form a height gradient distribution. A local coordinate system is established to calculate the normal and tangential pressure components at each point on the occipital surface, determining the pressure distribution direction. Specifically, in the cervical lordosis region (usually the C4-C5 segment), the height difference between adjacent support zones should be designed so that the tangent of the occipital surface is tangent to the cervical spine's physiological curve. In the local coordinate system, using the tangential plane at each point on the occipital surface as a reference, the pressure acting on that point is decomposed into the normal component F. n and tangential component F t In ideal conditions, F n It should be greater than F t And F t The direction should be towards the bony support structure.

[0069] By raising the height of the corresponding zones of the cervical spine's bony support structures and lowering the height of the corresponding zones of the soft tissue distribution areas, a gradient field is created that guides the pressure distribution towards the bony support structures. Specifically, the height of the occipital support zone can be set as h1, the upper cervical support zone as h2, the middle cervical support zone as h3, and the lower cervical support zone as h4. To create an effective height gradient, h1 > h2 and h4 > h3. Typically, the height difference between h1 and h2 is maintained within the range of 10-15 mm, and the height difference between h4 and h3 is maintained within the range of 5-10 mm. Through this design, when gravity is applied to the head and neck, a gradient field is formed that transmits pressure from the soft tissue areas to the bony support structures, reducing the stress on the soft tissues and allowing the bony structures to bear more support.

[0070] The distribution ratio of cervical spine force under gradient field action within the bony support structure is calculated. When the distribution ratio reaches a preset force transmission threshold, the height value of each support zone is recorded. Assuming the total weight of the head and neck is W, the force borne by the bony support structure is measured as W using a pressure sensor array. b Then the force transmission ratio β = W b / W. When β reaches the preset threshold (usually 0.65-0.75, indicating that 65%-75% of the force is supported by the bony structure), record the current values ​​of h1, h2, h3, and h4. Substitute these height values ​​into the bivariate correlation function F(P, S) for verification. If F(P, S) ≤ K, then the set of height values ​​is confirmed to be the optimal height adjustment combination that satisfies the coupling constraint conditions.

[0071] In practical applications, an iterative optimization method can be used to adjust the height of each zone with an initial step size of 5mm, gradually reducing the precision to 1mm, and finally determining the precise height value of each support zone. To verify the adjustment effect, sleep monitoring technology can be used to evaluate the user's sleep quality indicators after using the optimized occipital structure, including sleep efficiency, sleep posture maintenance time, and subjective comfort score, and comprehensively evaluate the actual effect of the optimal height adjustment combination.

[0072] In practical applications, to improve user experience, multiple preset adjustment modes can be set, such as normal mode, cervical spine protection mode, and deep sleep mode. In different modes, the force transmission threshold and constraint weights will be adjusted. For example, in cervical spine protection mode, the force transmission ratio of the bony support structure will be strengthened to improve support stability; while deep sleep mode focuses more on even pressure distribution and reduces body movement frequency. Users can choose a suitable mode according to their individual needs and sleep patterns. The system will also continuously optimize the adjustment scheme based on the long-term accumulation of user sleep data, providing a more personalized sleep experience.

[0073] In one optional implementation, multiple independent support units within the pillow body are driven to perform differentiated height changes according to the height adjustment combination, forming a three-dimensional support surface adapted to the current head and neck posture, including: Based on the height adjustment combination, the adjustment difference between the target height and the current height of each support zone is calculated. When the height adjustment difference between adjacent support zones exceeds the preset sudden change limit, a transition height sequence is calculated based on linear interpolation. The transition sequence is then allocated to the independent support units between adjacent support zones to achieve continuous height transition. Based on the attitude stability index, the adjustment rate of each independent support unit is calculated, and the target height and adjustment rate are sent to each independent support unit. Pressure timing data is collected during the lifting and lowering process. The pressure change caused by the unit height change is calculated to obtain the pressure response coefficient. When the pressure response coefficient exceeds the pressure change threshold, the system returns to the previous height position and recalculates the adjustment rate. The process continues until the target height is reached. After each independent support unit reaches the target height, a three-dimensional support surface that adapts to the head and neck posture is formed.

[0074] In this specific embodiment, based on head and neck posture information and height adjustment combination data, the difference between the target height and the current height of each support zone is calculated. Whether height transition processing is needed is determined by comparing the height adjustment difference between adjacent support zones to see if it exceeds a preset abrupt change limit (e.g., 10mm). When the height adjustment difference between two adjacent zones exceeds this limit, a linear interpolation algorithm is used to calculate the transition height sequence.

[0075] The linear interpolation algorithm works based on the following formula: For a support unit i between adjacent partitions A and B, its transition height is: H(i) = H(A) + (H(B) - H(A)) × (iA) / (BA), where H(A) and H(B) are the target heights of partitions A and B, respectively, and i represents the unit position. For example, if the target height of partition A is 50mm and the target height of partition B is 70mm, and there are 4 support units in between, then the assigned transition height sequence is [54, 58, 62, 66]mm.

[0076] These transition sequences are assigned to independent support units between adjacent support zones to ensure a smooth height transition between different support areas and avoid discomfort to the head and neck from abrupt support surfaces.

[0077] The adjustment rate of each independent support unit is calculated based on the posture stability index. The posture stability index is determined by the uniformity of pressure distribution in the head and neck area and the dynamic change rate of the support area. For units supporting important areas of the head (such as the occipital lobe), a lower adjustment rate (such as 2-5 mm / s) is used to ensure stable support; for units in the neck transition area, a moderate rate (such as 5-8 mm / s) can be used; and for the edge auxiliary area, a higher rate (such as 8-10 mm / s) can be used.

[0078] The system sends target height and adjustment rate commands to each independent support unit. The drive unit then begins to perform height adjustment. The command includes the unit ID, target height value (accurate to 0.1 mm), and adjustment rate parameter (mm / s). After receiving the command, the support unit's internal drive mechanism (such as a micro stepper motor or linear actuator) performs lifting and lowering actions at the specified rate.

[0079] During the lifting and lowering process, the pressure sensor array built into the pillow body collects pressure time-series data in real time at a sampling frequency of 20Hz to capture the dynamic contact state between the head and neck and the support surface. By calculating the pressure change caused by a unit height change, the pressure response coefficient is obtained: PRC = ΔP / ΔH, where ΔP represents the pressure change value (Pascals) and ΔH represents the height change value (mm).

[0080] When the pressure response coefficient exceeds the preset pressure change threshold (e.g., 2.5 Pa / mm), the system identifies a sudden support state that may cause discomfort to the user. At this time, a height retraction operation is performed to retract the support unit to the previous height position (usually retracted by 2-3 mm), and a smoother adjustment rate is recalculated (generally reduced by 25%-50%).

[0081] For example, if a support unit is initially adjusted at a rate of 5 mm / s, and the pressure response coefficient is detected to reach 3 Pa / mm (exceeding the threshold), the unit will be retracted to the previous height position and the adjustment will continue at a rate of 2.5 mm / s. This dynamic adaptive adjustment process continues until all support units reach their respective target heights.

[0082] Once all the independent support units reach the target height, a three-dimensional support surface is formed that precisely matches the head and neck posture. This surface typically presents a support shape that is the opposite of the natural physiological curve of the cervical spine, forming a moderate depression in the occipital lobe area and a reasonable protrusion in the neck area, ensuring that the head and neck maintain their natural physiological curves when at rest, reducing cervical spine pressure and improving sleep quality.

[0083] In a practical application scenario, when a user transitions from a supine to a side-lying position, the system detects the change in head and neck posture and calculates a new target height configuration. During the adjustment process, the support units near the shoulder-neck junction require a greater range of height adjustment to adapt to the side-lying position. The system automatically calculates and assigns a smooth transition height sequence between these adjacent units. Simultaneously, it monitors pressure response in real time to ensure a smooth and comfortable adjustment process, ultimately forming an ideal three-dimensional support surface in the side-lying position, effectively relieving neck pressure points and achieving natural alignment of the head, neck, and spine.

[0084] The sleep posture monitoring and intelligent pillow height adjustment system of this invention includes: The data acquisition unit is used to acquire time-series data of head and neck contact pressure of the target user during the sleep cycle; The posture analysis and stability assessment unit is used to calculate the pressure distribution matrix based on the time series data of head and neck contact pressure, obtain the position of the pressure centroid and its migration trajectory, identify the head and neck posture transition mode by combining the morphological change characteristics of the pressure peak region, and calculate the posture stability index according to the spatial gradient and temporal evolution characteristics of the pressure distribution. The pressure distribution non-uniformity calculation unit is used to divide the sleep cycle into a posture transition stage and a posture stability stage according to the posture transition mode. For the posture stability stage, the pressure topology skeleton of the cervical spine support area is extracted, and its topology complexity, density characteristics and pressure propagation steepness are calculated. The pressure distribution non-uniformity index is calculated in combination with the pressure range. The cervical curvature parameter extraction unit is used to extract the personalized cervical physiological curvature parameters of the target user within the initial usage period. The support height optimization calculation unit is used to optimize the pressure distribution direction by adjusting the height gradient between adjacent support zones of the pillow body based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, combined with the cervical spine physiological curvature parameters, so that the force on the cervical spine is transferred to the bony support structure, and calculate the optimal height adjustment combination of different support zones of the pillow body. The dynamic adjustment execution unit is used to drive multiple independent support units within the pillow body to perform differentiated height changes according to the height adjustment combination, forming a three-dimensional support surface that adapts to the current head and neck posture.

[0085] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0086] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0087] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring sleep posture and intelligently adjusting pillow height, characterized in that, include: Acquire time-series data of head and neck contact pressure of the target user during the sleep cycle; Based on the time-series data of head and neck contact pressure, the pressure distribution matrix is ​​calculated to obtain the position of the pressure centroid and its migration trajectory. Combined with the morphological change characteristics of the pressure peak region, the head and neck posture transition mode is identified. At the same time, the posture stability index is calculated based on the spatial gradient and temporal evolution characteristics of the pressure distribution. Based on the posture transition pattern, the sleep cycle is divided into a posture transition phase and a posture stabilization phase. For the posture stabilization phase, the pressure topology skeleton of the cervical spine support area is extracted, and its topological complexity, density characteristics and pressure propagation steepness are calculated. The pressure distribution unevenness index is calculated in combination with the pressure range. During the initial usage period, the personalized cervical spine physiological curvature parameters of the target user are extracted; Based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, and combined with the cervical spine physiological curvature parameters, the pressure distribution direction is optimized by adjusting the height gradient between adjacent support zones of the pillow body, so that the force on the cervical spine is transferred to the bony support structure, and the optimal height adjustment combination of different support zones of the pillow body is calculated. The height adjustment combination drives multiple independent support units within the pillow to perform differentiated height changes, forming a three-dimensional support surface that adapts to the current head and neck posture.

2. The method according to claim 1, characterized in that, Based on the time-series data of head and neck contact pressure, a pressure distribution matrix is ​​calculated to obtain the location of the pressure centroid and its migration trajectory. Combined with the morphological changes in the pressure peak region, head and neck posture transition patterns are identified. Simultaneously, based on the spatial gradient and temporal evolution characteristics of the pressure distribution, posture stability indices are calculated, including: The time-series data of head and neck contact pressure is sampled and segmented at fixed time intervals. The position of the pressure centroid of the pressure distribution matrix in each time period is calculated, and the migration trajectory is obtained by connecting them in time sequence. The pressure values ​​in the pressure distribution matrix are converted into pressure potential energy per unit area. The pressure potential energy gradient between adjacent pressure sampling points is calculated. An energy flow vector field characterizing the pressure transmission path and intensity distribution is generated. The energy inflow-outflow flux ratio in the pressure peak region is calculated to obtain the energy accumulation rate. Extract the equivalent radius of the pressure peak region, calculate the correlation coefficient between the equivalent radius and the energy accumulation rate, and determine that a head and neck attitude transition has occurred when the correlation coefficient is negative and its absolute value exceeds the attitude determination threshold. Based on the migration trajectory and the principal direction of the energy flow vector field, determine the head and neck attitude transition mode. The gradient distribution of the pressure distribution matrix in the horizontal, vertical and diagonal directions is calculated, the standard deviation and extreme point density of the gradient in each direction are extracted as spatial feature parameters, the local change frequency of the pressure distribution matrix is ​​calculated, and the pressure time series data is decomposed into fast-changing components and slow-changing components. The spatial characteristic parameters are multiplied by the energy entropy of the rapidly changing component, and then differentially fused with the trend change of the slowly changing component to generate the attitude stability index.

3. The method according to claim 1, characterized in that, Based on the posture transition pattern, the sleep cycle is divided into a posture transition phase and a posture stabilization phase. For the posture stabilization phase, the pressure topology skeleton of the cervical spine support area is extracted, and its topological complexity, density characteristics, and pressure propagation steepness are calculated. Combined with the pressure range, the pressure distribution unevenness index is calculated, including: The time of attitude transition in the attitude transition mode is marked as the transition boundary point of the sleep cycle. The time integral value of the attitude stability index between adjacent transition boundary points is calculated. The time period when the integral value is higher than the stability integral threshold is marked as the attitude stability stage, and the rest is marked as the attitude transition stage. For the posture stabilization phase, the cervical spine support area is extracted and its pressure distribution is morphologically refined. A pressure topology skeleton with a single pixel width is extracted, the number of branch points is counted to obtain the topology complexity, and the ratio of the total skeleton length to the region area is calculated to obtain the density feature. The pressure topology skeleton is traced from the high pressure point to the low pressure point, the pressure drop rate is calculated, and the maximum pressure drop rate is used as the stress propagation steepness. The cervical spine support area is divided into grid cells, the pressure range within each grid cell is calculated and weighted summation is performed, and the pressure distribution non-uniformity index is obtained by combining the topological complexity, the density characteristics and the stress propagation steepness.

4. The method according to claim 3, characterized in that, For the posture stabilization phase, the cervical spine support region is extracted and its pressure distribution is morphologically refined. A single-pixel width pressure topology skeleton is extracted, the number of branch points is counted to obtain the topological complexity, and the density feature is obtained by calculating the ratio of the total skeleton length to the region area, including: Based on the pressure distribution data during the posture stabilization phase, a binary pressure image of the cervical spine contact area is segmented by setting a pressure threshold, which serves as the initial boundary of the cervical spine support area. Morphological thinning is performed on the binarized stress image. Under the constraints of maintaining regional connectivity and topology, the outer pixels are peeled off layer by layer. For each pixel to be deleted, connectivity is checked and endpoint protection is determined to avoid regional splitting or the generation of isolated pixels, until a continuous and complete single-pixel width stress topology skeleton is obtained. Traverse all pixels of the pressure topology skeleton, identify pixels with more than the critical number of adjacent pixels as branch points, count the total number of branch points to obtain the topology complexity, calculate the cumulative distance between adjacent pixels based on the skeleton pixel coordinate sequence to obtain the total skeleton length, count the number of all effective pixels in the cervical spine support area as the area, and calculate the ratio of the total skeleton length to the area to obtain the skeleton distribution density per unit area, which is used as the density feature.

5. The method according to claim 1, characterized in that, During the initial usage period, the personalized cervical spine physiological curvature parameters of the target user are extracted, including: During multiple posture stabilization phases in the initial use period, pressure distribution data of the cervical spine support area are collected, spatial interpolation is performed on the pressure distribution data to generate a continuous pressure field, isobar clusters of the continuous pressure field are calculated, and the closed contours and centroids of each isobar are extracted. The centroids of isobars under different pressure values ​​are connected from high to low pressure to form a centroid series curve. The point of maximum curvature is identified as the projection position of the cervical spine curvature vertex. The curve is divided into the occipital bone support segment and the cervical spine support segment. The average radius of curvature and segment length ratio of the two segments are calculated. The distribution of the average radius of curvature and the ratio of segment length within different sleep days was statistically analyzed. The median of each parameter was calculated, outliers with interquartile ranges exceeding the median tolerance were removed, and the remaining data were weighted and averaged to obtain the stable radius of curvature and the stable ratio of segment length. The standard curvature distribution of the cervical spine support segment is determined based on the stable curvature radius, and its range of action in the occipital support area is determined by the proportion of the stable segment length. The peak position of the curvature distribution is determined by the projection position of the cervical spine flexure vertex, thus obtaining the personalized cervical spine physiological curvature parameters of the target user.

6. The method according to claim 1, characterized in that, Based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, and combined with the cervical spine physiological curvature parameters, the pressure distribution direction is optimized by adjusting the height gradient between adjacent support zones of the pillow body, so that the force on the cervical spine is transferred to the bony support structure. The optimal height adjustment combination for different support zones of the pillow body is calculated, including: Using the pressure distribution non-uniformity index as a spatial constraint variable and the posture stability index as a time constraint variable, a bivariate correlation function characterizing the coupling constraint relationship is constructed; the curvature information of each segment in the cervical spine physiological curvature parameters is extracted and mapped to each support zone of the occipital body to determine the location of the cervical spine bony support structure and the soft tissue distribution area; Based on the cervical spine physiological curvature parameters, the target height difference between adjacent support zones is calculated to form a height gradient distribution. A local coordinate system is established to calculate the normal pressure component and tangential pressure component at each point on the pillow surface to determine the pressure distribution direction. Increase the height of the corresponding zone of the cervical vertebral bony support structure and decrease the height of the corresponding zone of the soft tissue distribution area to form a gradient field that guides the pressure distribution direction to be transmitted to the bony support structure; Calculate the distribution ratio of cervical spine force in the bony support structure under the action of the gradient field. When the distribution ratio reaches the preset force transmission threshold, record the height value of each support zone. Substitute the height value into the bivariate correlation function for verification. When the constraint conditions are met, determine it as the optimal height adjustment combination.

7. The method according to claim 1, characterized in that, The height adjustment combination drives multiple independent support units within the pillow body to perform differentiated height changes, forming a three-dimensional support surface adapted to the current head and neck posture, including: Based on the height adjustment combination, the adjustment difference between the target height and the current height of each support zone is calculated. When the height adjustment difference between adjacent support zones exceeds the preset sudden change limit, a transition height sequence is calculated based on linear interpolation. The transition sequence is then allocated to the independent support units between adjacent support zones to achieve continuous height transition. Based on the attitude stability index, the adjustment rate of each independent support unit is calculated, and the target height and adjustment rate are sent to each independent support unit. Pressure timing data is collected during the lifting and lowering process. The pressure change caused by the unit height change is calculated to obtain the pressure response coefficient. When the pressure response coefficient exceeds the pressure change threshold, the system returns to the previous height position and recalculates the adjustment rate. The process continues until the target height is reached. After each independent support unit reaches the target height, a three-dimensional support surface that adapts to the head and neck posture is formed.

8. A sleep posture monitoring and intelligent pillow height adjustment system, used to implement the method as described in any one of claims 1-7, characterized in that, include: The data acquisition unit is used to acquire time-series data of head and neck contact pressure of the target user during the sleep cycle; The posture analysis and stability assessment unit is used to calculate the pressure distribution matrix based on the time series data of head and neck contact pressure, obtain the position of the pressure centroid and its migration trajectory, identify the head and neck posture transition mode by combining the morphological change characteristics of the pressure peak region, and calculate the posture stability index according to the spatial gradient and temporal evolution characteristics of the pressure distribution. The pressure distribution non-uniformity calculation unit is used to divide the sleep cycle into a posture transition stage and a posture stability stage according to the posture transition mode. For the posture stability stage, the pressure topology skeleton of the cervical spine support area is extracted, and its topology complexity, density characteristics and pressure propagation steepness are calculated. The pressure distribution non-uniformity index is calculated in combination with the pressure range. The cervical curvature parameter extraction unit is used to extract the personalized cervical physiological curvature parameters of the target user within the initial usage period. The support height optimization calculation unit is used to optimize the pressure distribution direction by adjusting the height gradient between adjacent support zones of the pillow body based on the coupling constraint relationship between the pressure distribution non-uniformity index and the posture stability index, combined with the cervical spine physiological curvature parameters, so that the force on the cervical spine is transferred to the bony support structure, and calculate the optimal height adjustment combination of different support zones of the pillow body. The dynamic adjustment execution unit is used to drive multiple independent support units within the pillow body to perform differentiated height changes according to the height adjustment combination, forming a three-dimensional support surface that adapts to the current head and neck posture.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images