Personalized body supporting and adjusting method and device based on intelligent mattress

Abstract

The invention discloses a personalized body support adjustment method and device based on an intelligent mattress, belongs to the technical field of intelligent mattresses, and aims to solve the technical problem that an existing intelligent mattress system is difficult to realize accurate pressure distribution identification, personalized support strength matching and dynamic adjustment strategy optimization. The method comprises the following steps: identifying a real-time sleeping posture and a body part division result of a user, and obtaining air bag basic adjustment time and a state parameter value of each body part; when the real-time sleeping posture of the user changes, a real-time pressure distribution matrix of the user is collected through a pressure array sensor, noise filtering is carried out, and a denoised pressure distribution matrix is obtained; according to the de-noising pressure distribution matrix, performing weight detection on the user to obtain user weight data; according to the weight data of the user and the state parameter values, the air bag personalized adjustment time corresponding to each body part is determined, and the folding air bags corresponding to the body parts are driven to be subjected to inflation and deflation adjustment.

Description

Technical Field

[0001] The present invention relates to the technical field of smart mattresses, and in particular to a personalized body support adjustment method and device based on a smart mattress. Background Art

[0002] During sleep, some parts of the human body will enter a relaxed state, and some parts will be in an empty state without support, which may make the user feel uncomfortable during sleep, leading to frequent tossing and turning, decreased sleep quality and other problems.

[0003] Research shows that different sleeping positions lead to different pressure distributions on different parts of the body. While existing smart mattress systems can provide differentiated support for different parts of the user's body through local support adjustments, they have significant technical flaws:

[0004] First, the raw data collected by the pressure sensor contains complex noise components. Especially in dynamic scenarios such as when the user turns over, the signal-to-noise ratio drops significantly. During the signal processing process, the traditional threshold filtering algorithm tends to mix and filter out the physiological signals of low-pressure areas such as the legs with environmental noise, and lacks adaptive capabilities, resulting in incomplete extraction of pressure features in key areas such as the legs and waist.

[0005] Second, a dynamic mapping mechanism between user physical parameters and support strength was not established. Changes in pressure thresholds caused by weight differences were not incorporated into the control system, resulting in an inadequate match between support strength and user needs. The weight prediction model relied solely on overall pressure values, failing to fully utilize the topological characteristics of pressure distribution, resulting in a lack of dimensionality in characterizing user characteristics.

[0006] Third, the fixed-duration adjustment cycle ignores the hysteresis effect of human tissue viscoelasticity on pressure response, which can easily trigger misadjustments in the presence of noise, affecting sleep continuity. It cannot adapt to the pressure response characteristics of users of different body shapes, nor can it maintain adjustment accuracy in noisy environments.

[0007] These defects together make it difficult for existing smart mattress systems to achieve accurate pressure distribution identification, personalized support strength matching, and dynamic adjustment strategy optimization, becoming the core technical bottleneck restricting the improvement of the sleeping experience of smart mattresses. Summary of the Invention

[0008] The embodiments of the present invention provide a personalized body support adjustment method and device based on a smart mattress, which is used to solve the following technical problems: existing smart mattress systems are difficult to achieve accurate pressure distribution identification, personalized support strength matching and dynamic adjustment strategy optimization.

[0009] The embodiment of the present invention adopts the following technical solutions:

[0010] In one aspect, an embodiment of the present invention provides a personalized body support adjustment method based on a smart mattress, wherein the smart mattress includes at least: a folded airbag array mounted on a mattress base and a pressure array sensor mounted on the surface of each folded airbag. The method includes:

[0011] Based on the sleeping posture recognition model, the system identifies the user's real-time sleeping posture and body part division results, and obtains the basic airbag adjustment time and status parameter values for each body part;

[0012] When the user's real-time sleeping posture changes, the user's real-time pressure distribution matrix is collected by the pressure array sensor, and noise is filtered out of the real-time pressure distribution matrix to obtain a denoised pressure distribution matrix;

[0013] performing weight detection on the user according to the denoised pressure distribution matrix to obtain user weight data;

[0014] According to the user weight data and the state parameter value, the personalized adjustment time of the airbag corresponding to each body part is determined, and according to the personalized adjustment time of the airbag and the basic adjustment time of the airbag, the folding airbag of the corresponding body part is driven to adjust the inflation and deflation.

[0015] In a feasible implementation, the pressure array sensor collects initial pressure distribution matrices of different users and different sleeping postures, and threshold filtering is performed on the pressure data in each initial pressure distribution matrix to obtain a set of denoised pressure distribution matrices;

[0016] generating a corresponding pressure distribution map according to each denoised pressure distribution matrix, marking key body parts in the pressure distribution map with rectangular boxes, and generating a regional mask matrix for each key body part;

[0017] Performing data enhancement on the body region elements and the corresponding region mask matrix in the denoised pressure distribution matrix, constructing a model training set, and training a preset sleeping posture recognition model;

[0018] The obtained real-time pressure distribution matrix is input into the sleeping posture recognition model to obtain the corresponding user's real-time sleeping posture and region mask matrix, and the corresponding body part division result is obtained according to the region mask matrix.

[0019] In a feasible implementation, obtaining the basic airbag adjustment time and state parameter values of each body part specifically includes:

[0020] Pre-set exclusive adjustment templates for different sleeping positions, and based on expert experience, set corresponding basic airbag adjustment times for different sleeping positions and associate them with the corresponding exclusive adjustment templates;

[0021] The state parameter values corresponding to the various body parts are set in the exclusive adjustment template; wherein the various body parts include at least shoulders, back, waist, buttocks, thighs and calves; when the state parameter value is greater than 0, the corresponding folding airbag inflation operation is performed; when the state parameter value is less than 0, the folding airbag deflation operation is performed; when the state parameter value is equal to 0, the folding airbag maintains the current state; the initial value of the state parameter value is 0.

[0022] In a feasible implementation, performing noise filtering on the real-time pressure distribution matrix to obtain a denoised pressure distribution matrix specifically includes:

[0023] For the i-th row data matrix in the real-time pressure distribution matrix i ,according to Calculate the corresponding row filter threshold Threshold i ; Where 1≤i≤m,1≤j≤n, m and n are the total number of rows and columns of the real-time pressure distribution matrix respectively; matrix i,j Represents the element in the i-th row and j-th column of the matrix matrix; α i is a hyperparameter, and 0<α i ≤1, used to adjust the i-th row data matrix i The filtering strength of α i The value of satisfies the condition: α of the hip i ≥α of the back i ≥α of shoulder i ≥ α at waist i , α of the thigh i >α of the calf i ;

[0024] The real-time pressure distribution matrix is filtered row by row using the row filtering threshold to obtain the denoised pressure distribution matrix.

[0025] In a feasible implementation, filtering the real-time pressure distribution matrix row by row using the row filtering threshold to obtain the denoised pressure distribution matrix specifically includes:

[0026] For the i-th row of the real-time pressure distribution matrix, traverse each element matrix in turn i,j ;

[0027] Each element matrix i,j The row filter threshold corresponding to the i-th row data i For comparison, if matrix i,j ≥Threshold i , then keep the element;

[0028] If matrix i,j <Threshold i , then set the element to 0;

[0029] Each row of filtered data is reconstructed into the denoised pressure distribution matrix.

[0030] In a feasible implementation, performing weight detection on the user according to the denoised pressure distribution matrix to obtain user weight data specifically includes:

[0031] According to w1=F w1 (sum(matrix_T)), determine the first estimated weight value w1 of the user; where matrix_T is the denoised pressure distribution matrix, sum(matrix_T) is the sum of all elements in matrix_T, F w1 is the preset first weight estimation function, which is a customized one-dimensional linear regression equation;

[0032] According to w2=F w2 (matrix_flattene), determine the second estimated weight value w2 of the user; wherein matrix_flattene is the data after the denoised pressure distribution matrix matrix_T is flattened into one dimension, F w2 The preset second weight estimation function is a customized multi-layer MLP model;

[0033] According to w=α*w1+β*w2, the user weight data w of the user is obtained; wherein α+β=1.

[0034] In a feasible implementation, determining the personalized airbag adjustment time corresponding to each body part according to the user weight data and the state parameter value specifically includes:

[0035] According to t ft =F w2t (w), obtain the airbag fine-tuning time t corresponding to each body part ft ; Wherein, w is the user's weight data, function F w2t (w) is the mapping equation between body weight and adjustment time. As w increases, F w2t The growth rate of (w) gradually decreases; t ft =(t1, t2, t3, t4, t5, t6), represents the time for fine-tuning the basic airbag adjustment time;

[0036] The airbag fine-tuning time t ftPerform bitwise addition with the corresponding state parameter value to obtain the personalized airbag adjustment time corresponding to each body part.

[0037] In a feasible implementation, according to the personalized airbag adjustment time and the basic airbag adjustment time, driving the foldable airbag corresponding to the body part to perform inflation and deflation adjustment specifically includes:

[0038] Determining, based on the personalized airbag adjustment time, an adjustment operation to be performed on the folding airbag in the area corresponding to the current body part; wherein the adjustment operation is a deflation operation, an inflation operation, or maintaining the airbag unchanged;

[0039] Subtracting the personalized airbag adjustment time from the basic airbag adjustment time bit by bit to obtain the actual airbag adjustment time of the folding airbag for each body part;

[0040] According to the adjustment operation and the actual adjustment time of the airbag, the folding airbag in the corresponding area is driven to perform inflation and deflation adjustment.

[0041] In a feasible implementation, after performing bitwise subtraction of the personalized airbag adjustment time from the basic airbag adjustment time to obtain the actual airbag adjustment time of the folding airbag for each body part, the method further includes:

[0042] The actual airbag adjustment time at this time is recorded and assigned to the basic airbag adjustment time for the next airbag adjustment calculation. This cycle is repeated so that each airbag adjustment is calculated based on the previous adjustment.

[0043] On the other hand, an embodiment of the present invention also provides a personalized body support adjustment device based on a smart mattress, the device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, so that the at least one processor can execute the personalized body support adjustment method based on the smart mattress.

[0044] Compared with the prior art, the personalized body support adjustment method and device based on the smart mattress provided by the embodiments of the present invention have the following beneficial effects:

[0045] 1. This invention utilizes an advanced sleeping posture recognition model and pressure array sensor to accurately identify the user's real-time sleeping posture and body part segmentation. Furthermore, an efficient row-level noise filtering algorithm removes complex noise components from the raw pressure data, ensuring a complete and accurate denoised pressure distribution matrix. This process effectively overcomes the incomplete pressure feature extraction caused by noise interference in traditional technologies, significantly improving the accuracy of pressure distribution recognition.

[0046] 2. This invention not only considers the user's real-time pressure distribution but also innovatively incorporates the user's weight data as a key parameter for adjusting support strength. By combining user weight data with preset state parameter values, the system dynamically determines the personalized airbag adjustment time corresponding to each body part, accurately matching support strength with user needs. This mechanism effectively solves the problem of insufficient matching of support strength with user needs in the existing technology, which is caused by the lack of a dynamic mapping mechanism between user vital signs and support strength.

[0047] 3. This invention abandons the traditional fixed-duration adjustment cycle and instead adopts a dynamic adjustment strategy based on the user's weight and body part characteristics. By calculating the personalized adjustment time for the airbag, the system can flexibly adapt to the pressure response characteristics of users of different body types, ensuring high-precision adjustment even in noisy environments. Furthermore, this strategy effectively avoids misadjustments caused by the hysteresis effect of human tissue viscoelasticity on pressure response, significantly improving sleep continuity and comfort.

[0048] 4. Combining the above technical advantages, this invention can significantly enhance the sleeping experience of a smart mattress. Through precise pressure distribution identification, personalized support strength matching, and dynamic adjustment strategy optimization, the system can provide users with a support environment more tailored to their physical needs, effectively alleviating sleep discomfort caused by improper sleeping posture or insufficient support, and providing users with a more adaptive sleep experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments described in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0050] Figure 1 A flow chart of a personalized body support adjustment method based on a smart mattress provided by an embodiment of the present invention;

[0051] Figure 2 A comparison diagram of support adjustment effects in a supine position provided by an embodiment of the present invention;

[0052] Figure 3 A comparison diagram of support adjustment effects in a side-lying posture provided by an embodiment of the present invention;

[0053] Figure 4 A schematic structural diagram of a personalized body support adjustment device based on a smart mattress provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0055] An embodiment of the present invention provides a personalized body support adjustment method based on a smart mattress. The method is applied to a smart mattress. A folding airbag array is installed on the base of the smart mattress, and a pressure array sensor is installed on the surface of the folding airbag array.

[0056] like Figure 1 As shown, the personalized body support adjustment method based on the smart mattress specifically includes steps S101-S104:

[0057] S101. Based on a sleeping posture recognition model, identify the user's real-time sleeping posture and body part division results, and obtain the basic airbag adjustment time and state parameter value of each body part.

[0058] Specifically, the pressure array sensor collects initial pressure distribution matrices for different users and sleeping positions. The pressure data in each initial pressure distribution matrix is threshold filtered to obtain a set of denoised pressure distribution matrices. A corresponding pressure distribution map is then generated based on each denoised pressure distribution matrix. Key body parts are marked in the pressure distribution map using rectangular boxes, and a regional mask matrix is generated for each key body part.

[0059] Furthermore, data augmentation is performed on the body region elements and corresponding region mask matrices in the denoised pressure distribution matrix to construct a model training set, and a preset sleeping posture recognition model is trained. The real-time pressure distribution matrix acquired during actual use of the smart mattress is then input into the sleeping posture recognition model to obtain the corresponding real-time sleeping posture and region mask matrix. The corresponding body part segmentation results are then obtained based on the region mask matrix.

[0060] Furthermore, exclusive adjustment templates corresponding to different sleeping positions are pre-established, and corresponding basic airbag adjustment times are established for different sleeping positions based on expert experience and associated with the corresponding exclusive adjustment templates.

[0061] Set the status parameter values corresponding to each body part in the exclusive adjustment template; wherein each body part includes at least shoulders, back, waist, buttocks, thighs and calves; when the status parameter value is greater than 0, the corresponding folding airbag inflation operation is performed; when the status parameter value is less than 0, the folding airbag deflation operation is performed; when the status parameter value is equal to 0, the folding airbag maintains the current state; the initial value of the status parameter value is 0.

[0062] As a viable implementation, differentiated adjustment solutions are built based on different sleeping positions, such as supine, prone, left, and right. Dedicated adjustment templates are developed for each sleeping position. Each template strictly adheres to ergonomic principles, achieving precise control of the airbags to meet the physiological needs and mechanical characteristics of the human body in different sleeping positions. This provides scientific and reasonable support for the shoulders, back, waist, hips, thighs, and calves.

[0063] Each adjustment template is associated with six state parameter values, denoted as t_1, t_2, t_3, t_4, t_5, and t_6, corresponding to the adjustment states of the shoulders, back, waist, hips, thighs, and calves, respectively. When t>0, the airbag inflates; when t<0, it deflates; and when t=0, the airbag maintains its current state, neither inflating nor deflating. This allows for precise adjustment of different body parts.

[0064] Different sleeping positions have different adjustment times. Let P∈{supine, prone, left side, right side}. Each sleeping position P corresponds to a basic airbag adjustment time t=(t1,t2,t3,t4,t5,t6), where t and P satisfy t=F(P). F returns the basic airbag adjustment time for each body part when the user is in sleeping position P.

[0065] S102 : When the user's real-time sleeping posture changes, the user's real-time pressure distribution matrix is collected through the pressure array sensor, and noise is filtered out of the real-time pressure distribution matrix to obtain a denoised pressure distribution matrix.

[0066] When a person lies on a mattress, the pressure on the leg area is relatively low. At the same time, the use of a smart mattress can generate significant noise interference. Traditional threshold filtering schemes have significant drawbacks, easily filtering out both the effective leg pressure information and the noise. To mitigate this noise interference and prevent leg information from being mistakenly filtered, the present invention employs a line filtering approach for noise removal.

[0067] Specifically, for the i-th row data matrix in the real-time pressure distribution matrix i ,according to Calculate the corresponding row filter threshold Threshold i; Where 1≤i≤m,1≤j≤n, m and n are the total number of rows and columns of the real-time pressure distribution matrix respectively; matrix i,j Represents the element in the i-th row and j-th column of the matrix matrix; α i is a hyperparameter, and 0<α i ≤1, used to adjust the i-th row data matrix i The filtering strength of α i The value of satisfies the condition: α of the hip i ≥α of the back i ≥α of shoulder i ≥ α at waist i , α of the thigh i >α of the calf i .

[0068] Furthermore, the real-time pressure distribution matrix is filtered row by row through the row filtering threshold to obtain the denoised pressure distribution matrix, which specifically includes:

[0069] For the i-th row of the real-time pressure distribution matrix, traverse each element matrix in turn i,j . Each element matrix i,j The row filter threshold corresponding to the i-th row data i For comparison, if matrix i,j ≥Threshold i , then keep the element. If matrix i,j <Threshold i , then set the element to 0. Reconstruct each row of filtered data into a denoised pressure distribution matrix.

[0070] As a feasible implementation, the present invention independently calculates a threshold for each row of pressure array data and then uses this threshold to filter each row of data. This process effectively filters out most of the noise generated by the human body and retains the most effective pressure information in the leg area.

[0071] S103: Detect the user's weight based on the denoised pressure distribution matrix to obtain the user's weight data.

[0072] Specifically, since the user's weight is proportional to the overall pressure value, weight prediction can be performed using the overall pressure value and each value in the pressure array. To reduce the impact of noise on weight prediction, we use the filtered, denoised pressure distribution matrix for dual weight prediction, which greatly improves the accuracy of weight prediction. The specific implementation process is as follows:

[0073] According to w1=F w1(sum(matrix_T)) determines the user's first weight estimate w1; where matrix_T is the denoised pressure distribution matrix, sum(matrix_T) is the sum of all elements in matrix_T, and F w1 is the preset first weight estimation function, which is a customized one-dimensional linear regression equation.

[0074] According to w2=F w2 (matrix_flattene), determine the user's second weight estimate w2; where matrix_flattene is the data after the denoised pressure distribution matrix matrix_T is flattened into one dimension, F w2 It is the preset second weight estimation function, which is a customized multi-layer MLP model.

[0075] Furthermore, the user weight data w of the user is obtained according to w=α*w1+β*w2; wherein α+β=1.

[0076] As a feasible implementation method, although the overall pressure value is positively correlated with the user's weight, it is not a strict linear fitting relationship, but is scattered on both sides of the positive proportional curve. If the weight prediction is performed only according to the strategy of the overall pressure value, it is very likely to produce a large error. In order to improve the prediction accuracy, in the process of weight prediction, in addition to the basis of the sum of the pressure values, the present invention also flattens the filtered matrix_T and performs multi-layer data perception. By comprehensively considering these factors, we strive to predict the user's weight more accurately, thereby further optimizing the effect of the airbag adjustment solution. By taking a weighted sum of these two weight values, the accuracy of weight prediction can be improved. Among them, F w1 is a one-dimensional linear regression equation. w2 It is a multi-layer MLP neural network model. In order to adapt to the input of MLP, matrix_T is first flattened into one dimension and spliced according to the rows. If matrix_T is an m*n matrix, then after flattening, matrix_flatten = (x 11 ,x 12 ,…,x m,n-1 ,x m,n ); where x ij Represents the element at row i and column j in matrix_T.

[0077] S104. Determine the personalized adjustment time of the airbag corresponding to each body part based on the user's weight data and status parameter values, and drive the folding airbag of the corresponding body part to perform inflation and deflation adjustment based on the personalized adjustment time and the basic adjustment time of the airbag.

[0078] Specifically, as the pressure on the smart mattress increases step by step, it will deform due to the greater pressure, and the greater the weight, the greater the deformation of the mattress. However, it is worth noting that as the deformation continues to increase, the impact of weight on the mattress is not constant, but rather shows a trend of gradually weakening. For example, when the weight increases from 50 kg to 60 kg, the mattress may have a more obvious change in sinking; but when the weight increases from 90 kg to 100 kg, the change in the degree of mattress sinking is relatively less significant compared to the same increase in weight before.

[0079] In view of the above characteristics of the smart mattress, the setting of its adjustment time needs to be reasonably planned. Therefore, when setting the adjustment time, the present invention consciously slows down the growth rate of the adjustment time as the user's weight increases. Specifically, the function t ft =F w2t (w) Mapping of weight and adjustment time, where F w2t (w) is a function whose increasing rate gradually decreases as w increases. The specific implementation process is as follows:

[0080] First, according to t ft =F w2t (w), obtain the airbag fine-tuning time t corresponding to each body part ft ;Where, w is the user's weight data, function F w2t (w) is the mapping equation between body weight and adjustment time. As w increases, F w2t The growth rate of (w) gradually decreases; t ft =(t1′, t2′, t3′, t4′, t5′, t6′), which represents the time for fine-tuning the basic adjustment time of the airbag.

[0081] Furthermore, according to t now =(t1′+t_1,t2′+t_2,t3′+t_3,t4′+t_4,t5′+

[0082] t_5,t6′+t_6), fine-tune the airbag time t ft The bitwise summation with the corresponding state parameter value is performed to obtain the personalized airbag adjustment time t corresponding to each body part now .

[0083] Furthermore, according to the size of the personalized airbag adjustment time, the adjustment operation to be performed by the folding airbag in the corresponding area of the current body part is determined. Specifically, when t now >0, corresponding to the airbag inflation operation; when t now <0, execute the airbag deflation operation; when t now=0, it means that the airbag maintains its current state, neither inflating nor deflating, thereby achieving fine adjustment of different body parts.

[0084] Furthermore, the personalized airbag adjustment time is subtracted from the basic airbag adjustment time bit by bit to obtain the actual airbag adjustment time for each body part. The actual airbag adjustment time is recorded and assigned to the basic airbag adjustment time for the next airbag adjustment calculation. This cycle is repeated so that each airbag adjustment is calculated based on the previous adjustment.

[0085] As a feasible implementation method, after obtaining the user's personalized adjustment time, corresponding operations need to be performed on the airbag. However, the current airbag is already in the basic adjustment state. For example, when the personalized adjustment time of the hip airbag is calculated to be 3 seconds of inflation, if the basic adjustment time of the hip airbag is also 3 seconds of inflation, in this case, the airbag does not need to perform any additional operations. For another example, if the personalized adjustment time of the hip airbag is calculated to be 3 seconds of inflation, and the current hip airbag has actually been deflated for 2 seconds based on the basic adjustment time - 2s, then in order to achieve the personalized adjustment state, the current hip airbag needs to be inflated for 5 seconds.

[0086] Based on this, the time of personalized airbag adjustment is obtained, and the basic airbag adjustment time of the current airbag is also determined. By calculating the difference between these two data, that is, the time corresponding to the personalized airbag adjustment time minus the time value corresponding to the basic airbag adjustment time, the time required for the current airbag operation can be accurately obtained. The specific formula is state = t now -t old , where t now Represents the personalized adjustment time of the airbag; t old It represents the basic airbag adjustment time. now With t old The difference operation is performed, and the result is the current operation time required to perform the airbag operation, which is recorded as state.

[0087] Whenever the airbag state is updated, the state value at that time is assigned to t old In this way, in the process of adjusting in sequence, t old The previous airbag status is always preserved.

[0088] Finally, according to the adjustment operation and the actual adjustment time of the airbag, the folding airbag in the corresponding area is driven to adjust the inflation and deflation.

[0089] As a feasible implementation method, in order to verify the adjustment effect, the present invention conducts a comparative analysis of the pressure distribution thermodynamic diagrams before and after the adjustment. Figure 2This is a comparison diagram of the support adjustment effect in a supine position provided by an embodiment of the present invention. Figure 3 This is a comparison diagram of the support adjustment effect under a side-lying posture provided by an embodiment of the present invention. The pressure distribution heat map is drawn based on the pressure value generated when the user lies on the bed. Figure 2 and Figure 3 By comparison, it is found that the pressure distribution heat map after adjustment is more uniform, which intuitively shows that the pressure distribution after support adjustment is more uniform, which means that every part of the user's body can get good support.

[0090] Furthermore, building on existing pressure array sensors, this invention can integrate temperature, humidity, and bioelectric sensors into the smart mattress. This technology automatically adjusts the mattress's heating and ventilation functions based on the user's body temperature and humidity, enhancing sleep comfort. By monitoring physiological indicators such as the user's heart rate and respiratory rate in real time, it provides data support for sleep quality assessment and health management.

[0091] It can also use machine learning algorithms to analyze the user's historical data on sleeping posture changes and establish a sleeping posture prediction model. This allows the airbag's inflation and deflation to be adjusted in advance before the user's sleeping posture is about to change, reducing discomfort caused by sleeping posture changes and improving sleep continuity.

[0092] Furthermore, the present invention can continuously optimize the support adjustment algorithm according to the user's sleeping habits to achieve more accurate personalized support adjustment.

[0093] Finally, based on the user's daily sleeping posture data and support adjustment data, a daily / weekly sleep quality report is generated, including data such as sleep duration, deep sleep ratio, number of tossing and turning, etc. Users can manually adjust the support strength and area of the mattress according to their preferences.

[0094] When a smart mattress is used by multiple people, it uses biometric recognition to distinguish between different users and saves individual sleep preferences for each user. When a user change is detected, the corresponding personalized support adjustment plan is automatically loaded. Multiple family members can share the same mattress, each enjoying a customized sleep experience.

[0095] In addition, the embodiment of the present invention also provides a personalized body support adjustment device based on a smart mattress, such as Figure 4 As shown, the personalized body support adjustment device based on the smart mattress specifically includes:

[0096] at least one processor; and a memory communicatively connected to the at least one processor; wherein,

[0097] The memory stores instructions executable by at least one processor, so as to enable the at least one processor to perform:

[0098] Based on the sleeping posture recognition model, the system identifies the user's real-time sleeping posture and body part division results, and obtains the basic airbag adjustment time and status parameter values for each body part;

[0099] When the user's real-time sleeping posture changes, the user's real-time pressure distribution matrix is collected by the pressure array sensor, and noise is filtered out of the real-time pressure distribution matrix to obtain a denoised pressure distribution matrix;

[0100] performing weight detection on the user according to the denoised pressure distribution matrix to obtain user weight data;

[0101] According to the user weight data and the state parameter value, the personalized adjustment time of the airbag corresponding to each body part is determined, and according to the personalized adjustment time of the airbag and the basic adjustment time of the airbag, the folding airbag of the corresponding body part is driven to adjust the inflation and deflation.

[0102] The various embodiments of the present invention are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the device, apparatus, and non-volatile computer storage medium embodiments are generally similar to the method embodiments, so their descriptions are simplified. For relevant details, refer to the descriptions of the method embodiments.

[0103] The above description of specific embodiments of the present invention is provided. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0104] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations may be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A personalized body support adjustment method based on a smart mattress, characterized in that: The smart mattress comprises at least: a folding airbag array mounted on a mattress base and a pressure array sensor mounted on the surface of each folding airbag. The method comprises: Based on the sleeping posture recognition model, the system identifies the user's real-time sleeping posture and body part division results, and obtains the basic airbag adjustment time and status parameter values for each body part; When the user's real-time sleeping posture changes, the user's real-time pressure distribution matrix is collected by the pressure array sensor, and noise is filtered out of the real-time pressure distribution matrix to obtain a denoised pressure distribution matrix; performing weight detection on the user according to the denoised pressure distribution matrix to obtain user weight data; According to the user weight data and the state parameter value, the personalized adjustment time of the airbag corresponding to each body part is determined, and according to the personalized adjustment time of the airbag and the basic adjustment time of the airbag, the folding airbag of the corresponding body part is driven to adjust the inflation and deflation.

2. The personalized body support adjustment method based on a smart mattress according to claim 1, characterized in that: Based on the sleeping posture recognition model, the user's real-time sleeping posture and body part segmentation results are identified, including: The pressure array sensor collects initial pressure distribution matrices of different users and different sleeping postures, and performs threshold filtering on the pressure data in each initial pressure distribution matrix to obtain a set of denoised pressure distribution matrices; generating a corresponding pressure distribution map according to each denoised pressure distribution matrix, marking key body parts in the pressure distribution map with rectangular boxes, and generating a regional mask matrix for each key body part; Performing data enhancement on the body region elements and the corresponding region mask matrix in the denoised pressure distribution matrix, constructing a model training set, and training a preset sleeping posture recognition model; The obtained real-time pressure distribution matrix is input into the sleeping posture recognition model to obtain the corresponding user's real-time sleeping posture and region mask matrix, and the corresponding body part division result is obtained according to the region mask matrix.

3. The personalized body support adjustment method based on a smart mattress according to claim 1, characterized in that: Obtain the basic airbag adjustment time and status parameter values for each body part, including: Pre-set exclusive adjustment templates for different sleeping positions, and based on expert experience, set corresponding basic airbag adjustment times for different sleeping positions and associate them with the corresponding exclusive adjustment templates; The state parameter values corresponding to the various body parts are set in the exclusive adjustment template; wherein the various body parts include at least shoulders, back, waist, buttocks, thighs and calves; when the state parameter value is greater than 0, the corresponding folding airbag inflation operation is performed; when the state parameter value is less than 0, the folding airbag deflation operation is performed; when the state parameter value is equal to 0, the folding airbag maintains the current state; the initial value of the state parameter value is 0.

4. The personalized body support adjustment method based on a smart mattress according to claim 1, characterized in that: Noise filtering is performed on the real-time pressure distribution matrix to obtain a denoised pressure distribution matrix, specifically including: For the i-th row data matrix in the real-time pressure distribution matrix i ,according to Calculate the corresponding row filter threshold Threshold i ; Where 1≤i≤m,1≤j≤n, m and n are the total number of rows and columns of the real-time pressure distribution matrix respectively; matrix i,j Represents the element in the i-th row and j-th column of the matrix matrix; α i is a hyperparameter, and 0<α i ≤1, used to adjust the i-th row data matrix i The filtering strength of α i The value of satisfies the condition: α of the hip i ≥α of the back i ≥α of shoulder i ≥ α at waist i , α of the thigh i >α of the calf i ; The real-time pressure distribution matrix is filtered row by row using the row filtering threshold to obtain the denoised pressure distribution matrix.

5. The personalized body support adjustment method based on a smart mattress according to claim 4, characterized in that: The real-time pressure distribution matrix is filtered row by row using the row filtering threshold to obtain the denoised pressure distribution matrix, specifically including: For the i-th row of the real-time pressure distribution matrix, traverse each element matrix in turn i,j ; Each element matrix i,j The row filter threshold corresponding to the i-th row of data i For comparison, if matrix i,j ≥Threshold i , then keep the element; If matrix i,j <Threshold i , then set the element to 0; Each row of filtered data is reconstructed into the denoised pressure distribution matrix.

6. The personalized body support adjustment method based on a smart mattress according to claim 1, characterized in that: Performing a weight test on the user according to the denoised pressure distribution matrix to obtain user weight data specifically includes: According to w1=F w1 (sum(matrix_T)), determine the first estimated weight value w1 of the user; where matrix_T is the denoised pressure distribution matrix, sum(matrix_T) is the sum of all elements in matrix_T, F w1 is the preset first weight estimation function, which is a customized one-dimensional linear regression equation; According to w2=F w2 (matrix_flattene), determine the second estimated weight value w2 of the user; wherein matrix_flattene is the data after the denoised pressure distribution matrix matrix_T is flattened into one dimension, F w2 The preset second weight estimation function is a customized multi-layer MLP model; According to w=α*w1+β*w2, the user weight data w of the user is obtained; wherein α+β=1.

7. The personalized body support adjustment method based on a smart mattress according to claim 1, characterized in that: Determining the personalized airbag adjustment time corresponding to each body part according to the user weight data and the state parameter value specifically includes: According to t ft =F w2t (w), obtain the airbag fine-tuning time t corresponding to each body part ft ; Wherein, w is the user's weight data, function F w2t (w) is the mapping equation between body weight and adjustment time. As w increases, F w2t The growth rate of (w) gradually decreases; t ft =(t1, t2, t3, t4, t5, t6), represents the time for fine-tuning the basic airbag adjustment time; The airbag fine-tuning time t ft Perform bitwise addition with the corresponding state parameter value to obtain the personalized airbag adjustment time corresponding to each body part.

8. The personalized body support adjustment method based on a smart mattress according to claim 1, characterized in that: According to the personalized airbag adjustment time and the basic airbag adjustment time, driving the folding airbag corresponding to the body part to perform inflation and deflation adjustment, specifically including: Determining, based on the personalized airbag adjustment time, an adjustment operation to be performed on the folding airbag in the area corresponding to the current body part; wherein the adjustment operation is a deflation operation, an inflation operation, or maintaining the airbag unchanged; Subtracting the personalized airbag adjustment time from the basic airbag adjustment time bit by bit to obtain the actual airbag adjustment time of the folding airbag for each body part; According to the adjustment operation and the actual adjustment time of the airbag, the folding airbag in the corresponding area is driven to perform inflation and deflation adjustment.

9. The personalized body support adjustment method based on a smart mattress according to claim 8, characterized in that: After performing bitwise subtraction of the personalized airbag adjustment time from the basic airbag adjustment time to obtain the actual airbag adjustment time of the folding airbag for each body part, the method further includes: The actual airbag adjustment time at this time is recorded and assigned to the basic airbag adjustment time for the next airbag adjustment calculation. This cycle is repeated so that each airbag adjustment is calculated based on the previous adjustment.

10. A personalized body support adjustment device based on a smart mattress, characterized in that: The device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, so that the at least one processor can execute the personalized body support adjustment method based on the smart mattress according to any one of claims 1 to 9.

Images