An adaptive mattress adjustment method and a smart mattress

Abstract

An adaptive mattress adjustment method and a smart mattress are disclosed, comprising acquiring a user's subjective state tags and objective body posture tags, generating a first mattress adjustment scheme based on the subjective state tags and objective body posture tags, and adjusting the mattress according to the first mattress adjustment scheme. The first mattress adjustment scheme includes an airbag to be adjusted, the airbag inflation / deflation action of the airbag to be adjusted, and the amount of air inflation / deflation. This addresses the technical problem that current adaptive mattress adjustment methods lack the ability to personalize for individual differences.

Description

Technical Field

[0001] This invention relates to the field of smart mattress technology, specifically to an adaptive mattress adjustment method and a smart mattress. Background Technology

[0002] With advancements in technology and improvements in living standards, people are paying increasing attention to sleep health. Smart mattresses, as a key means of improving sleep quality, have experienced rapid development in recent years. Smart mattresses can improve sleep quality by dynamically adjusting the firmness of the mattress.

[0003] Existing adaptive mattress adjustment methods primarily rely on data analysis and learning technologies to construct a highly intelligent sleep system. This system can collect vast amounts of data in real time from the sleep platform, including but not limited to local pressure information changing over time at multiple locations. This data provides a solid foundation for intelligent mattress adjustment. Simultaneously, by recognizing the user's lying posture and spinal shape, the mattress can automatically adjust the inflation and deflation of each air chamber based on the user's actual lying position or spinal shape to achieve optimal support. For example, when the user is lying on their back, the system can identify and locate three key positions: the most prominent point of the thoracic spine, the most concave point of the lumbar spine, and the most prominent point of the buttocks. Based on the height relationship between these points, it intelligently adjusts the corresponding air spring modules to ensure proper support of the spine's natural curve, effectively relieving body pressure and improving sleep comfort.

[0004] While these adaptive mattress adjustment methods improve the user experience to some extent, there is still room for improvement. Although current systems can recognize a user's sleeping posture and spinal shape and adjust the mattress based on the user's pressure information, they lack the ability to personalize the mattress to individual differences. Summary of the Invention

[0005] The main technical problem this invention addresses is the lack of personalized customization capabilities for individual differences in current adaptive mattress adjustment methods.

[0006] According to a first aspect, one embodiment provides an adaptive mattress adjustment method, the mattress including a plurality of airbags, the airbags being used to inflate or deflate body parts of a user for support or relaxation; the adaptive mattress adjustment method includes:

[0007] Obtain the user's subjective state tags, wherein the subjective state tags include user preference tags and bodily perception tags;

[0008] Obtain the user's objective body posture tags, wherein the objective body posture tags are used to indicate the user's body posture;

[0009] A first mattress adjustment scheme is generated based on the subjective state label and the objective body posture label, wherein the first mattress adjustment scheme includes an airbag to be adjusted, the airbag inflation and deflation action of the airbag to be adjusted, and the amount of gas inflation and deflation.

[0010] The mattress is adjusted according to the first mattress adjustment scheme.

[0011] In some embodiments, generating a first mattress adjustment scheme based on the subjective state label and the objective body posture label includes:

[0012] Based on the subjective state labels, the objective body posture labels, and the expert system, a first mattress adjustment scheme is generated. The expert system includes a knowledge base and a reasoning engine. The knowledge base includes rules for judging airbag inflation and deflation actions and rules for controlling the amount of gas inflation and deflation.

[0013] In some embodiments, it also includes:

[0014] At preset time intervals, obtain the user's first satisfaction score and first pressure state score for each mattress adjustment within that time period.

[0015] The expert system is trained and updated using the user's first satisfaction score and first pressure state score after each mattress adjustment;

[0016] Alternatively, at preset time intervals, obtain the user's first satisfaction score, first pressure state score, and first sleep quality score for each mattress adjustment within that time period.

[0017] The expert system is trained and updated using the user's first satisfaction score, first pressure state score, and first sleep quality score after each mattress adjustment.

[0018] In some embodiments, training the expert system to update it using the user's first satisfaction score, first pressure state score, and first sleep quality score after each mattress adjustment includes:

[0019] Each of the first satisfaction score, the first stress state score, and the first sleep quality score is assigned a corresponding weight.

[0020] The user’s first final satisfaction with the mattress after each adjustment is calculated based on the first satisfaction score, the first stress state score, the first sleep quality score, and their respective weights.

[0021] The first final satisfaction levels are sorted to obtain the sorted first final satisfaction levels. The mattress adjustment schemes corresponding to the M highest first final satisfaction levels in the sorted first final satisfaction levels are selected as the first recommended mattress adjustment schemes. The expert system is then trained and updated based on the first recommended mattress adjustment schemes, where M is an integer not less than 1.

[0022] In some embodiments, the expert system has been pre-trained, and the pre-training method includes:

[0023] Obtain training state labels and training posture labels, wherein the training state labels include training preference labels and training perception labels;

[0024] The training state labels, training posture labels, and the knowledge base are used as input data for the inference engine, and multiple reference mattress adjustment schemes are obtained through the logical reasoning process of the inference engine.

[0025] The mattress is adjusted according to each of the reference mattress adjustment schemes, and the user's second satisfaction with the adjusted mattress is obtained;

[0026] After adjusting the mattress according to each of the reference mattress adjustment schemes, obtain the pressure value of the user using the adjusted mattress, and calculate the second pressure state score of the user using the adjusted mattress based on the pressure value;

[0027] For each of the reference mattress adjustment schemes, the user's second final satisfaction is calculated based on the second satisfaction and the second pressure state score, and the second final satisfaction is sorted to obtain the sorted second final satisfaction.

[0028] The N reference mattress adjustment schemes corresponding to the highest second final satisfaction levels among the sorted second final satisfaction levels are selected as the second recommended mattress adjustment schemes. The expert system is then optimized based on the second recommended mattress adjustment schemes to improve user satisfaction, resulting in a pre-trained expert system, where N is an integer not less than 1.

[0029] In some embodiments, calculating the user's second final satisfaction based on the second satisfaction level and the second stress score includes:

[0030] Obtain the second sleep quality score after the user uses the adjusted mattress;

[0031] Each of the second satisfaction score, the second stress state score, and the second sleep quality score is assigned a corresponding weight.

[0032] The user's second final satisfaction is calculated based on the second satisfaction score, the second stress score, the second sleep quality score, and their respective weights.

[0033] In some embodiments, the mattress includes multiple pressure points, each pressure point being equipped with a corresponding pressure sensor. The pressure sensor detects the pressure value at each pressure point. The pressure value exerted by the user on the mattress includes the user's pressure value at each pressure point. Calculating a second pressure state score based on the pressure values ​​after the user has adjusted the mattress includes:

[0034] Multiple pressure feature values ​​are generated based on the user's pressure value at each pressure point, wherein the pressure feature values ​​include the pressure distribution area and the sum of the pressure values;

[0035] The score corresponding to each pressure feature value is calculated based on the preset score evaluation rules, and a corresponding weight is assigned to each pressure feature value. The score corresponding to the pressure feature value is used to represent the user's comfort level under the pressure feature value.

[0036] The user's second pressure state score is calculated based on the weight of each pressure feature value, the score corresponding to each pressure feature value, and the pressure state score calculation formula.

[0037] In some embodiments, the reference mattress adjustment scheme includes a reference airbag to be adjusted, the airbag inflation / deflation action and gas inflation / deflation volume of the reference airbag, and the training state label, training posture label, and knowledge base are used as input data for the inference engine. Multiple reference mattress adjustment schemes are obtained through the logical reasoning process of the inference engine, including:

[0038] The user's body parts to be adjusted are determined based on the training status label and the training posture label, and the corresponding reference airbags to be adjusted are determined based on the body parts to be adjusted.

[0039] The training state label, training posture label, and knowledge base are used as input data for the inference engine, so that the inference engine determines the airbag inflation / deflation action of the reference airbag based on the training state label, the training posture label, and the airbag inflation / deflation action judgment rule, and determines the gas inflation / deflation amount of the reference airbag based on the training state label, the training posture label, and the gas inflation / deflation amount control rule.

[0040] In some embodiments, the airbag inflation / deflation action judgment rule includes multiple reference state labels and reference body posture labels, the weight of the inflation / deflation action corresponding to the reference state label, the inflation / deflation action marker value corresponding to the reference state label, the weight of the inflation / deflation action corresponding to the reference body posture label, and the inflation / deflation action marker value corresponding to the reference body posture label. The reference state labels include reference preference labels and reference perception labels. The weight of the inflation / deflation action is used to indicate the probability of an inflation or deflation action occurring. The inflation / deflation action marker value is used to indicate whether the airbag is about to inflate or deflate. The airbag inflation / deflation action judgment rule is as follows:

[0041] If there is a training state label that matches the reference state label but there is no training body label that matches the reference body label, then the airbag inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference state label.

[0042] If there is a training posture label that matches the reference posture label but there is no training state label that matches the reference state label, then the airbag inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference posture label.

[0043] If both a training state label and a training posture label that are consistent with the reference state label exist simultaneously, then the inflation / deflation action marker value of the reference airbag is calculated based on the weight of the inflation / deflation action corresponding to the reference state label, the inflation / deflation action marker value corresponding to the reference state label, the weight of the inflation / deflation action corresponding to the reference posture label, and the inflation / deflation action marker value corresponding to the reference posture label. The airbag inflation / deflation action of the reference airbag is then determined based on the inflation / deflation action marker value of the reference airbag.

[0044] According to a second aspect, one embodiment provides a smart mattress, the smart mattress including multiple airbags, the airbags being used to inflate and support or deflate and relax parts of the user's body, thereby adjusting the smart mattress; the smart mattress further includes:

[0045] The processor is used to execute the adaptive mattress adjustment method.

[0046] According to the adaptive mattress adjustment method and smart mattress of the above embodiments, since the user's subjective state label and objective body posture label are obtained respectively, when generating the first mattress adjustment scheme, the subjective state label and objective body posture label are combined to determine the airbag to be adjusted, the airbag inflation and deflation action and the gas inflation and deflation amount of the airbag to be adjusted, and the mattress is adjusted according to the first mattress adjustment scheme with the determined airbag to be adjusted, the airbag inflation and deflation action and the gas inflation and deflation amount. Therefore, the user's subjective state and objective state are taken into account when adjusting the mattress, thereby providing a customized mattress adjustment scheme for each user and improving the user experience. Attached Figure Description

[0047] Figure 1 This is a flowchart of the adaptive mattress adjustment method according to an embodiment of this application;

[0048] Figure 2 A flowchart illustrating one embodiment of obtaining a user's objective body posture tags;

[0049] Figure 3 A full-body image of a user performing a preset action, as shown in one embodiment;

[0050] Figure 4 One embodiment includes a full-body image of a user containing joint points at different body parts;

[0051] Figure 5 A flowchart illustrating a method for pre-training an expert system according to one embodiment;

[0052] Figure 6 This is a flowchart illustrating how training state labels, training posture labels, and a knowledge base are used as input data for an inference engine, and how multiple reference mattress adjustment schemes are obtained through the logical reasoning process of the inference engine, as one embodiment.

[0053] Figure 7 A table showing the airbag inflation / deflation action judgment rules for one embodiment;

[0054] Figure 8 A table showing the gas charge / discharge control rules for one embodiment;

[0055] Figure 9 A flowchart illustrating, in one embodiment, a second pressure state score calculated based on pressure values ​​using a user-adjusted mattress;

[0056] Figure 10 A flowchart illustrating, in one embodiment, a user's second final satisfaction level calculated based on a second satisfaction level and a second stress state score;

[0057] Figure 11 This is a flowchart illustrating the iterative optimization process of an expert system according to one embodiment. Detailed Implementation

[0058] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of this application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to this application are not shown or described in the specification. This is to avoid obscuring the core parts of this application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0059] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0060] The serial numbers assigned to components in this document, such as "first" and "second," are used only to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this application, unless otherwise specified, include both direct and indirect connections (linkages).

[0061] Please refer to Figure 1 This invention provides an adaptive mattress adjustment method. The mattress includes multiple airbags, which are used to inflate and support or deflate and relax the user's body parts. The adaptive mattress adjustment method includes steps S10 to S40, which are described in detail below.

[0062] Step S10: Obtain the user's subjective state label.

[0063] In some embodiments, the user's subjective state label includes user preference label and body perception label. The user preference label includes the user's preference label for the mattress, which may be the user's preference label for the firmness of the mattress or the user's preference label for the height of the pillow. The body perception label includes the user's perception label for strain on different parts of their own body, which may be a label for shoulder strain or a label for back strain.

[0064] In some embodiments, user preference tags and body perception tags can be obtained through questionnaires.

[0065] Step S20: Obtain the user's objective body posture labels.

[0066] In some embodiments, objective posture labels are used to indicate the user's posture, such as forward head posture, anterior pelvic tilt, and bowlegs.

[0067] Please refer to Figure 2 In some embodiments, step S20 obtains the user's objective body shape label, including steps S21 to S22, which are described in detail below.

[0068] Step S21: Obtain a full-body image of the user when performing a preset action, and identify the joints at different parts of the user's body based on the full-body image.

[0069] Please refer to Figure 3 In this embodiment, a full-body image of the user standing normally is obtained from the front, left, back, and right sides, and a full-body image of the user leaning back is obtained from the right side.

[0070] Please refer to Figure 4 Based on the acquired full-body image, the system identifies the joints at different parts of the user's body and marks them with red dots.

[0071] Step S22: Determine the user's objective posture label based on the joints at different body parts and the preset posture judgment rules.

[0072] In this embodiment, the joints at different body parts include the joints at the user's neck, shoulders, femur, knees, and feet. The angles between the joints at different body parts are calculated, and the user's objective posture label is determined based on the calculated angles and posture judgment rules. For example, the angle between the joints at the neck and the joints at the shoulders is calculated, and whether the angle is greater than a certain preset threshold in the posture judgment rules is used to determine whether the user has the objective posture label of forward head posture.

[0073] In some embodiments, objective body shape labels of users can be obtained through a body measurement mirror or a 3D scanner.

[0074] Step S30: Generate the first mattress adjustment scheme based on subjective state labels and objective body posture labels.

[0075] In this embodiment, the first mattress adjustment scheme includes an airbag to be adjusted, the airbag inflation / deflation action of the airbag to be adjusted, and the gas inflation / deflation amount. The user's body part to be adjusted is determined based on subjective state tags and objective body posture tags. The corresponding airbag to be adjusted is determined based on the body part to be adjusted, and the airbag inflation / deflation action and gas inflation / deflation amount of the airbag to be adjusted are further determined.

[0076] In some embodiments, a first mattress adjustment scheme is generated based on subjective state labels and objective body posture labels, including:

[0077] The first mattress adjustment plan is generated based on subjective state labels, objective body posture labels, and an expert system.

[0078] In some embodiments, the expert system includes a knowledge base and an inference engine. The knowledge base includes rules for judging airbag inflation / deflation actions and rules for controlling gas inflation / deflation volume. The knowledge base is derived from experts with backgrounds in ergonomics, sports medicine, and related fields. The inference engine can perform logical reasoning based on the rules in the knowledge base. It receives user input or tags and questions generated internally by the system, and then uses the rules and information in the knowledge base to perform step-by-step reasoning. By matching rules, applying premises, and deriving conclusions, it gradually solves the user's problem or achieves the system's goal.

[0079] In some embodiments, a first mattress adjustment scheme is generated based on subjective state labels, objective body posture labels, and a trained scheme generation model. The scheme generation model is trained using machine learning techniques and can generate a corresponding mattress adjustment scheme based on the input labels.

[0080] Please refer to Figure 5 In some embodiments, the expert system has been pre-trained. The pre-training method includes steps S31 to S36, which are described in detail below.

[0081] Step S31: Obtain training status labels and training posture labels.

[0082] In this embodiment, the training state label includes a training preference label and a training awareness label.

[0083] Step S32: Use the training state labels, training posture labels, and knowledge base as input data for the inference engine, and obtain multiple reference mattress adjustment schemes through the logical reasoning process of the inference engine.

[0084] In some embodiments, the reference mattress adjustment scheme includes a reference airbag to be adjusted, the airbag inflation / deflation action of the reference airbag, and the amount of gas inflation / deflation. Please refer to [reference needed]. Figure 6 Step S32 uses the training state label, training posture label and knowledge base as input data for the inference engine. Multiple reference mattress adjustment schemes are obtained through the logical reasoning process of the inference engine, including steps S321 to S322, which are explained in detail below.

[0085] Step S321: Determine the user's body parts to be adjusted based on the training status label and training posture label, and determine the corresponding reference airbags to be adjusted based on the body parts to be adjusted.

[0086] In this embodiment, the training status labels include training preference labels and training perception labels collected through subjective questionnaires. The training status labels include low pillow, preference for a firm mattress, no upper back strain, preference for relaxed upper back, significant lower back strain & acute phase, and preference for relaxed lower back. The training posture labels include standard back shape (i.e., no hunchback), severe anterior pelvic tilt, protruding buttocks, and small backward tilt angle. When the training status label is a combination of shoulder strain and preference for mattress firmness, the training status labels include significant shoulder strain ∩ preference for a firm mattress, significant shoulder strain ∩ preference for a soft mattress, mild shoulder strain ∩ preference for a firm mattress, and mild shoulder strain ∩ preference for a soft mattress. When the training posture labels are head & shoulder posture labels, the training posture labels include severe forward head posture ∩ no severe rounded shoulders / anterior humeral displacement, obvious forward head posture ∩ no severe rounded shoulders / anterior humeral displacement, and excessively small backward tilt angle. The training posture labels also include excessively small backward tilt angle, which indicates insufficient shoulder flexibility in the user. When the training status label is a combination of shoulder strain and preference for mattress firmness, and the training posture label is a head & shoulder posture label, both the training status label and the training posture label involve adjustments to the user's shoulders. Therefore, the user's body part to be adjusted is determined to be the shoulders. Since each body part of the user has a corresponding airbag, the corresponding reference airbag to be adjusted is determined based on the body part to be adjusted.

[0087] Step S322: Use the training state label, training posture label, and knowledge base as input data for the inference engine, so that the inference engine can determine the airbag inflation / deflation action of the reference airbag based on the training state label, training posture label, and airbag inflation / deflation action judgment rules, and determine the gas inflation / deflation amount of the reference airbag based on the training state label, training posture label, and gas inflation / deflation amount control rules.

[0088] In this embodiment, the airbag inflation / deflation action judgment rule includes multiple reference state labels and reference body posture labels, the weight of the inflation / deflation action corresponding to the reference state label, the inflation / deflation action marker value corresponding to the reference state label, the weight of the inflation / deflation action corresponding to the reference body posture label, and the inflation / deflation action marker value corresponding to the reference body posture label. The reference state labels include reference preference labels and reference perception labels. The weight of the inflation / deflation action indicates the probability of an inflation or deflation action occurring, and the inflation / deflation action marker value indicates that the airbag is about to inflate or deflate. The airbag inflation / deflation action judgment rule is as follows:

[0089] If there is a training state label that matches the reference state label but there is no training body label that matches the reference body label, then the airbag inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference state label.

[0090] If there is a training posture label that matches the reference posture label but there is no training state label that matches the reference state label, then the airbag inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference posture label.

[0091] If both a training state label and a training posture label that are consistent with the reference state label exist, then the inflation / deflation action label value of the reference airbag is calculated based on the weight of the inflation / deflation action corresponding to the reference state label, the label value of the inflation / deflation action corresponding to the reference state label, the weight of the inflation / deflation action corresponding to the reference posture label, and the label value of the inflation / deflation action corresponding to the reference posture label. The inflation / deflation action of the reference airbag is then determined based on the label value of the inflation / deflation action of the reference airbag.

[0092] In some embodiments, the inflation / deflation action marker value of the reference airbag is obtained by calculating the product of the weight of the inflation / deflation action corresponding to the reference status label and the inflation / deflation action marker value corresponding to the reference status label, and the product of the weight of the inflation / deflation action corresponding to the reference body posture label and the inflation / deflation action marker value corresponding to the reference body posture label, and summing the two products. When the inflation / deflation action marker value is greater than a preset judgment threshold, the inflation / deflation action of the reference airbag is determined to be inflation; otherwise, the inflation / deflation action of the reference airbag is determined to be deflation.

[0093] Please refer to Figure 7 The reference status labels are: significant shoulder strain (prefers a hard bed), significant shoulder strain (prefers a soft bed), mild shoulder strain (prefers a hard bed), and mild shoulder strain (prefers a soft bed). The reference posture labels are: severe forward head posture (no severe rounded shoulders / anterior humeral displacement), significant forward head posture (no severe rounded shoulders / anterior humeral displacement), and insufficient backward tilt angle. Specifically, the weight of the filling / releasing action corresponding to the reference status label "significant shoulder strain (prefers a hard bed)" is 5, and its corresponding filling / releasing action label value is 1. The weight of the filling / releasing action corresponding to the reference status label "significant shoulder strain (prefers a soft bed)" is 4, and its corresponding filling / releasing action label value is 1. The weight of the filling / releasing action corresponding to the reference posture label "severe forward head posture (no severe rounded shoulders / anterior humeral displacement)" is 2, and its corresponding filling / releasing action label value is -1. For details on the weights and filling / releasing action labels corresponding to other reference status labels or reference posture labels, please refer to [reference needed]. Figure 7 As shown in the figure. Among them, the inflation / deflation action marker value of 1 indicates that the airbag under this label is inflated, and the inflation / deflation action marker value of -1 indicates that the airbag under this label is deflated.

[0094] In some embodiments, when determining the inflation / deflation action of the reference airbag according to the airbag inflation / deflation action judgment rule, there are three cases, which are described in detail below. The first case: If there is a training state label consistent with the reference state label but no training posture label consistent with the reference posture label, then the inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference state label. For example, if there is a training state label corresponding to the reference state label "significant shoulder strain ∩ preference for a hard bed," and no training posture label consistent with the reference posture label, then the inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference state label. Figure 7 When the training posture label is consistent with any of the reference posture labels, the inflation / deflation action of the reference airbag is determined to be inflation based on the inflation / deflation action marker value 1 corresponding to the reference state label of "more significant shoulder strain ∩ preference for hard bed".

[0095] The second scenario: If a training posture label exists that matches the reference posture label but no training state label exists that matches the reference state label, then the inflation / deflation action of the reference airbag is determined based on the inflation / deflation action marker value corresponding to that reference posture label. For example, if a training posture label exists that corresponds to the reference posture label "severe head forward tilt ∩ no severe rounded shoulder / anterior humeral displacement," and no training state label exists that matches the reference state label, then the inflation / deflation action of the reference airbag is determined based on the inflation / deflation action marker value corresponding to that reference posture label. Figure 7 If any of the reference state labels matches the training state label, then the inflation / deflation action of the reference airbag is determined to be deflation based on the inflation / deflation action marker value -1 corresponding to the reference posture label of severe head forward tilt ∩ no severe rounded shoulders / anterior humeral tilt.

[0096] The third scenario: If both a training state label and a training posture label that match the reference state label exist simultaneously, then the inflation / deflation action marker value of the reference airbag is calculated based on the weights of the inflation / deflation actions corresponding to the reference state label, the marker values ​​of the inflation / deflation actions corresponding to the reference state label, the weights of the inflation / deflation actions corresponding to the reference posture label, and the marker values ​​of the inflation / deflation actions corresponding to the reference posture label. The inflation / deflation action of the reference airbag is then determined based on this marker value. For example, if there is a training state label corresponding to the reference state label "significant shoulder strain ∩ preference for hard beds," and a training posture label corresponding to the reference posture label "severe forward head posture ∩ no severe rounded shoulder / anterior humeral displacement," then the inflation / deflation action marker value of the reference airbag is calculated based on the inflation / deflation action marker value 1 (weight 5) corresponding to the reference state label "significant shoulder strain ∩ preference for hard beds," and the inflation / deflation action marker value -1 (weight 2) corresponding to the reference posture label "severe forward head posture ∩ no severe rounded shoulder / anterior humeral displacement." The product of the weight of the inflation / deflation action corresponding to the reference status label and the inflation / deflation action marker value corresponding to the reference status label is calculated as 5*1, and the product of the weight of the inflation / deflation action corresponding to the reference body posture label and the inflation / deflation action marker value corresponding to the reference body posture label is calculated as 2*(-1). The two products are then summed to obtain the inflation / deflation action marker value 3. Assuming that the preset judgment threshold is 0, when the inflation / deflation action marker value 3 is greater than the preset judgment threshold 0, the inflation / deflation action of the reference airbag is determined to be inflation.

[0097] In some embodiments, training state labels, training posture labels, and a knowledge base are used as input data to the inference engine, enabling the inference engine to determine the gas inflation / deflation volume of the reference airbag based on the training state labels, training posture labels, and gas inflation / deflation volume control rules. The gas inflation / deflation volume control rules can refer to... Figure 8 As shown, Figure 8 The gas inflation / deflation control rules include multiple reference status labels, the airbag inflation / deflation action corresponding to each reference status label, the gas inflation / deflation amount corresponding to each reference status label, multiple reference body posture labels, the airbag inflation / deflation action corresponding to each reference body posture label, and the gas inflation / deflation amount corresponding to each reference body posture label.

[0098] In some embodiments, the gas inflation / deflation volume of the reference airbag is determined by judging whether the training status label or training posture label of the reference airbag falls within the reference status label or reference posture label in the gas inflation / deflation volume control rules, and by combining the airbag inflation / deflation action of the reference airbag. For example, compared with Figure 8 According to the gas inflation / deflation control rules, when the reference status label of the reference airbag is "no shoulder strain" and the airbag inflation / deflation action is "inflation", the gas inflation / deflation amount is determined to be 20g. When the reference status label of the reference airbag is "no shoulder strain" and "prefers a hard bed", and the airbag inflation / deflation action is "deflation", the gas inflation / deflation amount is determined to be 40g.

[0099] In some embodiments, the gas inflation / deflation control rule may also include a combination of a reference state label and a reference body position label, i.e., reference state label ∩ reference body position label. The gas inflation / deflation amount of the reference airbag is determined by judging whether the training state label and training body position label of the reference airbag fall into the reference state label ∩ reference body position label at the same time, and by combining the airbag inflation / deflation action of the reference airbag.

[0100] Step S33: Adjust the mattress according to each reference mattress adjustment scheme and obtain the user's second satisfaction with the adjusted mattress.

[0101] In some embodiments, after adjusting the mattress according to each reference mattress adjustment scheme, a questionnaire can be used to collect users' subjective satisfaction with the adjusted mattress. This subjective satisfaction can include overall user satisfaction, satisfaction with different body parts, and user satisfaction with whether the mattress feels too soft or too firm. The users' subjective satisfaction with the adjusted mattress is converted into a corresponding score, where subjective satisfaction is directly proportional to its score; that is, the higher the subjective satisfaction, the higher the corresponding score. Based on the weight of each subjective satisfaction level and its corresponding score, the user's second satisfaction level with the adjusted mattress is calculated.

[0102] In some embodiments, the user's second satisfaction level after using the adjusted mattress is calculated using a second satisfaction calculation formula, wherein the first satisfaction calculation formula is:

[0103] F1=(w1*q1+w2*q2+...+wn*qn) / (w1+w2+...+wn)

[0104] Where F1 represents the second level of satisfaction, w1, w2, ..., wn represent the weights corresponding to each subjective level of satisfaction, and q1, q2, ..., qn represent the scores corresponding to each subjective level of satisfaction.

[0105] Step S34: Obtain the pressure value of the mattress when the user uses the adjusted mattress after adjusting the mattress according to each reference mattress adjustment scheme, and calculate the second pressure state score of the user using the adjusted mattress based on the pressure value.

[0106] In some embodiments, the mattress includes multiple pressure points, each with a corresponding pressure sensor. The pressure sensor detects the pressure value at each pressure point. The pressure value exerted by the user on the mattress includes the user's pressure value at each pressure point. Please refer to [reference needed]. Figure 9 Step S34 calculates the second pressure state score of the mattress after the user has adjusted it based on the pressure value, including steps S341 to S343, which are explained in detail below.

[0107] Step S341: Generate multiple pressure feature values ​​based on the user's pressure value at each pressure point.

[0108] In some embodiments, pressure characteristic values ​​are calculated based on the pressure values ​​detected by the pressure sensors at each pressure point. These pressure characteristic values ​​include the pressure distribution area and the sum of the pressure values. Additionally, pressure characteristic values ​​may also include length, width, average pressure value, kurtosis, bilateral symmetry, and a fitted body shape curve. The pressure distribution area refers to the total area of ​​the region where the user's body contacts the mattress and generates effective pressure, reflecting the actual contact between the user's body and the mattress. Length typically refers to the maximum longitudinal dimension of the contact area, and width refers to the maximum lateral dimension. Length and width together describe the two-dimensional distribution of the user's body on the mattress. The sum of pressure values ​​is the accumulation of all detected pressure values, reflecting the total pressure exerted by the user's body on the mattress and can be used to assess the mattress's support. The average pressure value is the arithmetic mean of all detected pressure values. Kurtosis is a statistical measure describing the shape of the pressure distribution, reflecting the sharpness of the pressure distribution. Bilateral symmetry assesses the left-right symmetry of the contact area between the user's body and the mattress. The fitted body shape curve is a contour curve of the user's body on the mattress, mathematically fitted using the pressure distribution detected by the sensors. The aforementioned pressure characteristics collectively constitute a comprehensive description of the pressure distribution when a user is on the mattress. By analyzing these pressure characteristics, one can gain a deeper understanding of the mattress's comfort and support.

[0109] Step S342: Calculate the score corresponding to each pressure feature value based on the preset score selection rules, and assign a corresponding weight to each pressure feature value.

[0110] In some embodiments, the preset scoring rules are rules pre-set by experts. For example, the scoring rules include assigning a high score to pressure characteristic values ​​with an average pressure value below 20 mmHg, and a lower score to higher average pressure values. This is used to calculate the score for each pressure characteristic value, and each pressure characteristic value is assigned a corresponding weight w11, w22, ..., wnn. The score corresponding to each pressure characteristic value represents the user's comfort level at that pressure characteristic value.

[0111] Step S343: Calculate the user's second pressure state score based on the weight of each pressure feature value, the score corresponding to each pressure feature value, and the pressure state score calculation formula.

[0112] In some embodiments, the formula for calculating the pressure state fraction is as follows:

[0113] F2=(w11*f1+w22*f2+...+wnn*fn) / (w11+w22+...+wnn)

[0114] Where F2 is the second pressure state score, f1, f2, ..., fn are the scores corresponding to each pressure feature value, and w11, w22, ..., wnn are the weights corresponding to each pressure feature value.

[0115] Step S35: For each reference mattress adjustment scheme, calculate the user's second final satisfaction based on the second satisfaction score and the second pressure state score, and sort the second final satisfaction scores to obtain the sorted second final satisfaction scores.

[0116] In some embodiments, corresponding weights are assigned to the second satisfaction score and the second stress score, respectively, and the user's final satisfaction is calculated based on the second satisfaction score, the second stress score, and their respective weights.

[0117] Please refer to Figure 10 In some embodiments, step S35 calculates the user's second final satisfaction based on the second satisfaction score and the second stress state score, including steps S351 to S353, which are described in detail below.

[0118] Step S351: Obtain the user's second sleep quality score after using the adjusted mattress.

[0119] In this embodiment, the second sleep quality score is a quantitative indicator used to measure the sleep quality obtained by a user using an adjusted mattress within a specific time period. Specifically, the user's sleep data, such as heart rate and respiratory rate, can be monitored in real time using sensors or other devices built into the mattress, and the second sleep quality score is calculated using an algorithm.

[0120] Step S352: Assign corresponding weights to the second satisfaction score, the second stress state score, and the second sleep quality score, respectively.

[0121] Step S353: Calculate the user's second final satisfaction based on the second satisfaction score, the second stress state score, the second sleep quality score, and their respective weights.

[0122] In some embodiments, the formula for calculating the user's second final satisfaction is as follows:

[0123] F=(W1*F1+W2*F2+W3*F3) / (W1+W2+W3)

[0124] Where F represents the second final satisfaction level, F1 represents the second satisfaction level, F2 represents the second stress level score, F3 represents the second sleep quality score, W1 represents the weight corresponding to the second satisfaction level, W2 represents the weight corresponding to the second stress level score, and W3 represents the weight corresponding to the second sleep quality score.

[0125] Step S36: Select the N reference mattress adjustment schemes with the highest second final satisfaction among the sorted second final satisfaction as the second recommended mattress adjustment schemes, and optimize the expert system according to the second recommended mattress adjustment schemes to improve user satisfaction, thus obtaining a pre-trained expert system.

[0126] In some embodiments, N is an integer not less than 1. The N reference mattress adjustment schemes corresponding to the highest second final satisfaction levels among the sorted second final satisfaction levels are selected as the second recommended mattress adjustment schemes. The inference engine in the expert system is optimized based on these second recommended mattress adjustment schemes, enabling the inference engine to learn the recommended mattress adjustment schemes. This allows the trained expert system to generate more satisfactory mattress adjustment schemes for users during subsequent scheme generation.

[0127] In other embodiments, the knowledge base is updated by obtaining user feedback on the reference mattress adjustment schemes corresponding to the N highest second final satisfaction levels, resulting in an updated knowledge base. The training state labels, training body posture labels, and the updated knowledge base are used as input data for the inference engine. The updated reference mattress adjustment schemes are obtained through the logical reasoning process of the inference engine until the user's satisfaction with the mattress adjusted according to the updated reference mattress adjustment scheme is greater than or equal to the satisfaction threshold. The inference engine and the updated knowledge base are used as a pre-trained expert system.

[0128] Please refer to Figure 11 In some embodiments, Figure 11 A method for training and updating an expert system is provided. The pre-trained expert system can generate multiple mattress adjustment schemes, and for each scheme, the user's first satisfaction score and first pressure state score are obtained. Based on these scores, a corresponding first final satisfaction score is derived, representing the first final satisfaction score for each mattress adjustment scheme. These final satisfaction scores are then ranked and input into an inference engine, enabling the engine to recommend a first mattress adjustment scheme. Throughout this process, real-time user feedback on the mattress is continuously collected and used to train the expert system, thus achieving iterative optimization.

[0129] In some embodiments, generating a first mattress adjustment scheme based on subjective state labels, objective body posture labels, and an expert system further includes steps S301 to S304, which are described in detail below.

[0130] Step S301: At each preset time interval, obtain the user's first satisfaction score and first pressure state score for each mattress adjustment within that time period.

[0131] In this embodiment, the preset time period can be several days or a week. Assuming the preset time period is a week, the user can adjust the mattress once or multiple times each day within that week. This adjustment can be based on the user's actual needs or on the user's initial satisfaction with the adjusted mattress; no limitation is made here. It is worth noting that the mattress adjustment can refer to the first mattress adjustment scheme.

[0132] In some embodiments, a questionnaire can be used to collect users' subjective satisfaction with the adjusted mattress. This subjective satisfaction can include overall user satisfaction, satisfaction with different body parts, and satisfaction with whether the mattress feels too soft or too firm. The user's subjective satisfaction with the adjusted mattress is then converted into a corresponding score, where the subjective satisfaction is directly proportional to the corresponding score; that is, the higher the subjective satisfaction, the higher the score. The user's initial satisfaction level with the adjusted mattress is calculated based on the weight of each subjective satisfaction level and its corresponding score.

[0133] In some embodiments, the first pressure state score of the user on the mattress after each adjustment within that time period can be obtained by referring to how the second pressure state score is calculated in step S34, which will not be repeated here.

[0134] Step S302: Use the user's first satisfaction score and first pressure state score for each mattress adjustment to train and update the expert system.

[0135] In some embodiments, by assigning corresponding weights to the first satisfaction score and the first pressure state score respectively, a user's satisfaction with the mattress after each adjustment is calculated based on the first satisfaction score, the first pressure state score and their respective weights. The satisfaction scores are then sorted, and the mattress adjustment schemes corresponding to the Q highest satisfaction scores in the sorted satisfaction scores are selected as mattress adjustment schemes to be recommended. The expert system is then trained and updated based on the mattress adjustment schemes to be recommended, where Q is an integer not less than 1.

[0136] Step S303: Alternatively, at preset time intervals, obtain the user's first satisfaction score, first pressure state score, and first sleep quality score for each mattress adjustment within that time period.

[0137] In this embodiment, the first sleep quality score is a quantitative indicator used to measure the sleep quality obtained by a user using an adjusted mattress within a specific time period. Specifically, the user's sleep data, such as heart rate and respiratory rate, can be monitored in real time using sensors or other devices built into the mattress, and the first sleep quality score is calculated using an algorithm.

[0138] Step S304: Use the user's first satisfaction score, first pressure state score, and first sleep quality score for each mattress adjustment to train and update the expert system.

[0139] In some embodiments, step S304 uses the user's first satisfaction with the mattress after each adjustment, the first pressure state score, and the first sleep quality score to train and update the expert system, including steps S304a to S304c, which are described in detail below.

[0140] Step S304a: Assign corresponding weights to the first satisfaction score, the first stress state score, and the first sleep quality score, respectively.

[0141] Step S304b: Calculate the user's first final satisfaction with the mattress after each adjustment based on the first satisfaction score, the first stress state score, the first sleep quality score, and their respective weights.

[0142] Step S304c: Sort the first final satisfaction levels to obtain the sorted first final satisfaction levels. Select the mattress adjustment schemes corresponding to the M highest first final satisfaction levels from the sorted first final satisfaction levels as the first recommended mattress adjustment schemes. Train the expert system based on the first recommended mattress adjustment schemes to update the expert system, where M is an integer not less than 1.

[0143] In this embodiment, the mattress adjustment scheme used by the user will be different each time the mattress is adjusted. After adjusting the mattress using the mattress adjustment scheme, the user's first satisfaction, first pressure state score and first sleep quality score of the adjusted mattress will also be different. After determining the M highest first final satisfactions among the sorted first final satisfactions, the mattress adjustment scheme corresponding to the first final satisfaction can be used as the first recommended mattress adjustment scheme, and the expert system can be updated according to the first recommended mattress adjustment scheme.

[0144] In some embodiments, the expert system is trained and updated using the user's first satisfaction score and first pressure state score for each mattress adjustment, or the expert system is trained and updated using the user's first satisfaction score, first pressure state score, and first sleep quality score for each mattress adjustment. Since the first satisfaction score, first pressure state score, and first sleep quality score are all obtained by the user through each mattress adjustment within a preset time period, these accumulated index values ​​of different dimensions can enable the expert system to combine the user's actual usage and feedback status during training, so that the updated expert system can recommend a more accurate and customized mattress generation solution.

[0145] Step S40: Adjust the mattress according to the first mattress adjustment plan.

[0146] Just as everyone's musculoskeletal condition and preferences are unique, mattresses should also be designed to provide individualized support and adjustment models. The adaptive mattress adjustment method described in this application is based on a thorough understanding of the body's needs, customizing professional and personalized adjustment plans according to the principles of ergonomics. For example, people with lumbar strain need adequate support for their lower back to provide additional stability and reduce muscle fatigue accumulated while standing or sitting. However, if in the acute phase of strain, the muscles and soft tissues are still inflamed and tense; excessive support can exacerbate local pressure, requiring appropriate pressure relief. For instance, when lying on one's side, the cervical spine needs to be positioned as close to a neutral position as possible; adjusting the pillow and shoulder airbags can reduce lateral bending of the cervical spine during sleep. Similarly, X-shaped legs can cause excessive stress on the inner side of the knee joint; the legs need appropriate pressure relief to reduce mattress pressure on the inner knee, allowing the knee to relax more naturally. This extreme personalization aims to achieve a more reasonable natural spinal alignment and body pressure distribution, reducing unnecessary exertion, promoting muscle rest and repair, and ultimately helping users achieve better energy recovery.

[0147] In some embodiments, assuming the user's posture is head-tilted, a trained expert system will suggest moderately deflating the shoulder and back airbags. When the user lies flat, moderately deflating the shoulder and back airbags can relax overly tense anterior muscle groups, especially the sternocleidomastoid and anterior, middle, and posterior scalene muscles. These muscles are constantly tense due to prolonged head tilting, leading to cervical lordosis. Simultaneously, deflating the shoulder airbags reduces the pressure on the scapulae, allowing them to naturally descend to a more suitable position. As the scapulae descend, the pulling force of the sternocleidomastoid and scalene muscles decreases, reducing excessive strain on the cervical spine. This process also activates and relaxes the posterior muscle groups, especially the upper trapezius and rhomboid muscles. Weakness in these posterior muscle groups is often the cause of scapular elevation; by deflating and adjusting their normal tension, the scapulae are restabilized on the rib cage, effectively improving the alignment of the scapular girdle with the cervical spine. The deflation of the back airbags further acts on the upper trapezius and latissimus dorsi muscles, helping them relieve excessive tension caused by abnormal scapular positioning, while also reducing pressure on the cervical spine. In practice, multiple adjustment options are provided based on different users' subjective preferences (such as mattress firmness) and objective conditions (such as weight, height, etc.) resulting in pressure distribution and body shape on the bed. Based on the user's subjectively set firmness (initial value) and objective conditions, different degrees of adjustment are made according to the actual pressure distribution and body shape on the mattress; that is, the inflation and deflation of the airbags are adjusted, and the most suitable option is determined through comfort indicators (including subjective satisfaction and objective indicators, such as pressure distribution ratio and body shape curve).

[0148] In some embodiments, the present invention provides a smart mattress, which includes multiple airbags for inflating or deflating body parts to support or relax the user, thereby adjusting the smart mattress. The smart mattress also includes a processor for executing an adaptive mattress adjustment method. The adaptive mattress adjustment method has been described in detail in steps S10 to S40 and will not be repeated here.

[0149] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0150] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. An adaptive mattress adjustment method, the mattress comprising a plurality of airbags, the airbags being used to inflate and support or deflate and relax parts of a user's body; characterized in that, The adaptive mattress adjustment method includes: Obtain the user's subjective state tags, wherein the subjective state tags include user preference tags and bodily perception tags; Obtain the user's objective body posture tags, wherein the objective body posture tags are used to indicate the user's body posture; A first mattress adjustment scheme is generated based on the subjective state label and the objective body posture label, wherein the first mattress adjustment scheme includes an airbag to be adjusted, the airbag inflation and deflation action of the airbag to be adjusted, and the amount of gas inflation and deflation. The mattress is adjusted according to the first mattress adjustment scheme.

2. The adaptive mattress adjustment method as described in claim 1, characterized in that, The process of generating a first mattress adjustment scheme based on the subjective state label and the objective body posture label includes: Based on the subjective state labels, the objective body posture labels, and the expert system, a first mattress adjustment scheme is generated. The expert system includes a knowledge base and a reasoning engine. The knowledge base includes rules for judging airbag inflation and deflation actions and rules for controlling the amount of gas inflation and deflation.

3. The adaptive mattress adjustment method as described in claim 2, characterized in that, Also includes: At preset time intervals, obtain the user's first satisfaction score and first pressure state score for each mattress adjustment within that time period. The expert system is trained and updated using the user's first satisfaction score and first pressure state score after each mattress adjustment; Alternatively, at preset time intervals, obtain the user's first satisfaction score, first pressure state score, and first sleep quality score for each mattress adjustment within that time period. The expert system is trained and updated using the user's first satisfaction score, first pressure state score, and first sleep quality score after each mattress adjustment.

4. The adaptive mattress adjustment method as described in claim 3, characterized in that, The process of training and updating the expert system using the user's first satisfaction score, first pressure state score, and first sleep quality score after each mattress adjustment includes: Each of the first satisfaction score, the first stress state score, and the first sleep quality score is assigned a corresponding weight. The user’s first final satisfaction with the mattress after each adjustment is calculated based on the first satisfaction score, the first stress state score, the first sleep quality score, and their respective weights. The first final satisfaction levels are sorted to obtain the sorted first final satisfaction levels. The mattress adjustment schemes corresponding to the M highest first final satisfaction levels in the sorted first final satisfaction levels are selected as the first recommended mattress adjustment schemes. The expert system is then trained and updated based on the first recommended mattress adjustment schemes, where M is an integer not less than 1.

5. The adaptive mattress adjustment method as described in claim 2, characterized in that, The expert system has undergone pre-training, and the pre-training methods include: Obtain training state labels and training posture labels, wherein the training state labels include training preference labels and training perception labels; The training state labels, training posture labels, and the knowledge base are used as input data for the inference engine, and multiple reference mattress adjustment schemes are obtained through the logical reasoning process of the inference engine. The mattress is adjusted according to each of the reference mattress adjustment schemes, and the user's second satisfaction with the adjusted mattress is obtained; After adjusting the mattress according to each of the reference mattress adjustment schemes, obtain the pressure value of the user using the adjusted mattress, and calculate the second pressure state score of the user using the adjusted mattress based on the pressure value; For each of the reference mattress adjustment schemes, the user's second final satisfaction is calculated based on the second satisfaction and the second pressure state score, and the second final satisfaction is sorted to obtain the sorted second final satisfaction. The N reference mattress adjustment schemes corresponding to the highest second final satisfaction levels among the sorted second final satisfaction levels are selected as the second recommended mattress adjustment schemes. The expert system is then optimized based on the second recommended mattress adjustment schemes to improve user satisfaction, resulting in a pre-trained expert system, where N is an integer not less than 1.

6. The adaptive mattress adjustment method as described in claim 5, characterized in that, The calculation of the user's second final satisfaction based on the second satisfaction level and the second stress level score includes: Obtain the second sleep quality score after the user uses the adjusted mattress; Each of the second satisfaction score, the second stress state score, and the second sleep quality score is assigned a corresponding weight. The user's second final satisfaction is calculated based on the second satisfaction score, the second stress score, the second sleep quality score, and their respective weights.

7. The adaptive mattress adjustment method as described in claim 5, wherein the mattress includes multiple pressure points, each pressure point is equipped with a corresponding pressure sensor, the pressure sensor is used to detect the pressure value at each pressure point, and the pressure value of the user on the mattress includes the pressure value of the user at each pressure point, characterized in that... The calculation of the second pressure state score of the mattress after user adjustment based on the pressure value includes: Multiple pressure feature values ​​are generated based on the user's pressure value at each pressure point, wherein the pressure feature values ​​include the pressure distribution area and the sum of the pressure values; The score corresponding to each pressure feature value is calculated based on the preset score evaluation rules, and a corresponding weight is assigned to each pressure feature value. The score corresponding to the pressure feature value is used to represent the user's comfort level under the pressure feature value. The user's second pressure state score is calculated based on the weight of each pressure feature value, the score corresponding to each pressure feature value, and the pressure state score calculation formula.

8. The adaptive mattress adjustment method as described in claim 5, characterized in that, The reference mattress adjustment scheme includes a reference airbag to be adjusted, the airbag inflation / deflation action and gas inflation / deflation volume of the reference airbag, and the training state label, training posture label, and knowledge base as input data to the inference engine. Multiple reference mattress adjustment schemes are obtained through the logical reasoning process of the inference engine, including: The user's body parts to be adjusted are determined based on the training status label and the training posture label, and the corresponding reference airbags to be adjusted are determined based on the body parts to be adjusted. The training state label, training posture label, and knowledge base are used as input data for the inference engine, so that the inference engine determines the airbag inflation / deflation action of the reference airbag based on the training state label, the training posture label, and the airbag inflation / deflation action judgment rule, and determines the gas inflation / deflation amount of the reference airbag based on the training state label, the training posture label, and the gas inflation / deflation amount control rule.

9. The adaptive mattress adjustment method as described in claim 8, characterized in that, The airbag inflation / deflation action judgment rule includes multiple reference state labels and reference body posture labels, the weight of the inflation / deflation action corresponding to the reference state label, the inflation / deflation action marker value corresponding to the reference state label, the weight of the inflation / deflation action corresponding to the reference body posture label, and the inflation / deflation action marker value corresponding to the reference body posture label. The reference state labels include reference preference labels and reference perception labels. The weight of the inflation / deflation action indicates the probability of an inflation or deflation action occurring. The inflation / deflation action marker value indicates that the airbag is about to inflate or deflate. The airbag inflation / deflation action judgment rule is as follows: If there is a training state label that matches the reference state label but there is no training body label that matches the reference body label, then the airbag inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference state label. If there is a training posture label that matches the reference posture label but there is no training state label that matches the reference state label, then the airbag inflation / deflation action of the reference airbag is determined according to the inflation / deflation action marker value corresponding to the reference posture label. If both a training state label and a training posture label that are consistent with the reference state label exist simultaneously, then the inflation / deflation action marker value of the reference airbag is calculated based on the weight of the inflation / deflation action corresponding to the reference state label, the inflation / deflation action marker value corresponding to the reference state label, the weight of the inflation / deflation action corresponding to the reference posture label, and the inflation / deflation action marker value corresponding to the reference posture label. The airbag inflation / deflation action of the reference airbag is then determined based on the inflation / deflation action marker value of the reference airbag.

10. A smart mattress, characterized in that, The smart mattress includes multiple airbags, which are used to inflate and support or deflate to relax parts of the user's body, thereby adjusting the smart mattress. The smart mattress also includes: A processor for performing the adaptive mattress adjustment method as described in any one of claims 1-9.

Images