Smart system and method for recommending a bed system

Abstract

The invention relates to a smart system for making personalised recommendations for bed systems and associated method. The system comprises a mattress with air chambers and pressure sensors, a BCG sensor, a user interface and a processing means connected to a reference database. The method collects anthropometric and physiological data of the user, analyses changes in pressure and determines the curvature of the back, comparing this data with stored profiles using a nearest-neighbour-based artificial intelligence algorithm. The system recommends the most suitable bed system to maximise the user's sleep quality.

Description

[0001] DESCRIPTION

[0002] INTELLIGENT SYSTEM AND PRESCRIPTION PROCEDURE FOR A REST SET

[0003] TECHNICAL SECTOR

[0004] The present invention relates to the field of sleep systems, especially to an intelligent sleep set prescription system and a computer-implemented sleep set prescription procedure.

[0005] BACKGROUND OF THE INVENTION

[0006] Sleep quality significantly impacts overall health and well-being. However, current recommendations for sleep systems often lack a thorough understanding of individual needs and are based solely on short-term trials in a commercial setting or on subjective buyer preferences.

[0007] There are some procedures for recommending mattresses based on generic databases that use BMI (Body Mass Index). These procedures rely on general assumptions, such as that a person with a high BMI requires a firmer mattress and a person with a low BMI requires a softer mattress.

[0008] More complex procedures also exist that map the pressure a person exerts on a test bed. One type of these procedures uses pressure blankets or sensors distributed across the bed's surface to obtain a two-dimensional representation of the user's body pressure distribution in one or more resting positions. A second type of procedure uses systems with air chambers or other sensor elements integrated into the test bed's structure itself, including both comfort and support layers. Based on these measurements, these procedures recommend a mattress that is supposed to reduce pressure points, for example, by selecting softer or firmer materials in the comfort layer or through specific mattress configurations.

[0009] However, known procedures have limitations that prevent the precise and reproducible prescription of a sleep system adapted to a specific user. In particular, although some of these procedures take into account the resting posture and perform measurements in more than one position, the measurements obtained primarily provide information regarding pressure distribution, without allowing for the establishment of a reliable quantitative relationship between these measurements and the overall mechanical characteristics of the sleep system.

[0010] Furthermore, these procedures do not allow for the objective determination of a mattress firmness parameter that reflects the combined behavior of the comfort and support layers, nor do they systematically integrate other relevant user parameters, such as physical condition, specific health conditions, or back shape. As a result, the recommendations obtained are often based on empirical or heuristic criteria and do not guarantee optimal adaptation to the individual needs of the user.

[0011] Therefore, there is a need for an improved procedure for prescribing a sleep set; more specifically, a procedure that allows for the objective and reproducible correlation of different user parameters with quantifiable technical characteristics of the sleep set, in order to determine a sleep set capable of providing better sleep quality to a particular person.

[0012] EXPLANATION OF THE INVENTION

[0013] The present invention relates to an intelligent system and method for determining and recommending a sleep set to a particular user, which contributes to a considerable improvement in their sleep quality. In this document, "sleep set" refers to a mattress, a pillow, a bed base, or a combination thereof.

[0014] Specifically, the object of the present invention, according to a first aspect, is an intelligent system for prescribing a sleep set. The system comprises:

[0015] a. A mattress comprising:

[0016] i. A control box with one or more pressure sensors,

[0017] i. A BCG sensor (ballistocardiography sensor) connected to the control box, iii. Air chambers connected to the control box,

[0018] with the pressure sensor or sensors configured to measure the pressure in the air chambers;

[0019] b. A user interface configured to input user data;

[0020] c. A set of reference data corresponding to respective reference users, comprising one or more of the following data types:

[0021] i. Sleep quality metrics,

[0022] i. Reference user profiles,

[0023] iii. Health data, iv. Temperature sensitivity,

[0024] v. Characteristics of the rest set used;

[0025] d. Data on prescriptive rest sets;

[0026] e. A programmable processing means configured to carry out the prescription of the rest set from the data obtained from steps a), b), c) and d).

[0027] Another object of the present invention according to a second aspect is a computer-implemented prescription procedure for a rest set, by means of an intelligent prescription system, comprising the following actions:

[0028] a. User data is entered through a user interface selected from the group comprising one or more of the following data types:

[0029] i. Anthropometric data,

[0030] i. Sleep habit data,

[0031] iii. Data related to health problems,

[0032] iv. Sensitivity to temperature;

[0033] b. The user lies down on a mattress comprising a control box with one or more pressure sensors, a BCG sensor (ballistocardiography sensor) connected to the control box, and air chambers connected to the control box, the pressure sensor(s) being configured to measure the pressure in the air chambers;

[0034] c. The pressure variation caused by the user is determined using the following steps:

[0035] i. The user is positioned supine;

[0036] i. A baseline measurement is taken in each chamber with the user at rest; iii. The air chambers are inflated simultaneously for a set time, preferably for 1 or 2 seconds;

[0037] iv. Another pressure measurement is taken in each chamber;

[0038] v. Steps (ii)-(iv) are repeated with the user in a lateral position;

[0039] vi. The pressure increase for each chamber is calculated for both postures; d. A session load (S) is calculated as an average value of the pressure increases measured in the air chambers in the supine and lateral postures;

[0040] e. The BCG sensor measures the user's vital signs;

[0041] f. The curvature of the user's back is determined using the following steps:

[0042] i. The user is positioned supine;

[0043] i. The air chambers are inflated for a certain time, sequentially in a longitudinal direction along the user's back; iii. The pressure reached in each chamber is measured, which depends on the resistance exerted by each area of ​​the person's body;

[0044] iv. The pressure values ​​achieved are transformed into shape values ​​according to a correspondence relationship between pressure and shape, the shape values ​​forming a curve that represents the curvature of the back; g. Using an AI proximity algorithm, preferably a k nearest neighbors (KNN) algorithm or similar, the data obtained in the previous stages a), c), d), e) and f) are compared with a set of reference data corresponding to respective reference users, said reference data comprising one or more of the following types of reference user data:

[0045] i. Sleep quality metrics,

[0046] i. Reference user profiles,

[0047] iii. Health data,

[0048] iv. Sensitivity to temperature,

[0049] v. Characteristics of the rest set used,

[0050] and the user is assigned the nearest reference user;

[0051] h. The rest set is prescribed to the user by assigning the rest set used by the nearest reference user.

[0052] In this document, the word "comprises" and its variants are to be interpreted as open-ended expressions that do not preclude the possibility of other technical features or components beyond those explicitly mentioned. Furthermore, the word "comprises" includes the case "consists of," which is interpreted as a closed-ended expression limited solely to the technical features or components explicitly mentioned. For those skilled in the art, other objects, advantages, and features of the invention will become apparent partly from the description and partly from the practice of the invention. Moreover, the present invention covers all possible combinations of embodiments described herein.

[0053] DESCRIPTION OF THE DRAWINGS

[0054] To complement the description being made and in order to help a better understanding of the characteristics of the invention, a set of drawings is included in which, for illustrative and non-limiting purposes, the following has been represented:

[0055] Figure 1: Data processing scheme.

[0056] Figure 2: Represents a sequence of air chamber inflation to determine the curvature of the back.

[0057] PREFERRED EMBODIMENT OF THE INVENTION

[0058] According to a first aspect, the present invention relates to a system for prescribing a personalized sleep set. A sleep set is understood to include at least a mattress, optionally a pillow or a base, or a combination thereof.

[0059] The system comprises four main elements:

[0060] A) An in-store trial set, consisting of a smart mattress, and optionally a smart pillow and adjustable base, said mattress uses advanced technologies including a BCG sensor (ballistocardiography sensor) and air chambers, both connected to a control box with one or more pressure sensors, as well as a user interface for entering customer information such as age, sex, temperature sensitivity and / or known health problems.

[0061] B) A system for classifying mattresses, pillows and bases (or a database of those already classified) according to their physical characteristics, mainly the level of comfort (for example, according to ISO 23769, it is firmness level Hs for mattresses, or according to ISO 2439 for pillows), but also thermal conductivity, number of zones, shape (in the case of a pillow), number of movements and characteristics (in the case of a base, for example anti-Trendelenburg movement), etc.

[0062] C) A database with historical data of real people consisting of quantified sleep quality, the characteristics of the sleep set used by this person, the physical characteristics of that person and their health status, their sensitivity to temperature, etc.

[0063] D) Software based on an AI algorithm (typically the k nearest neighbors (KNN) algorithm) that uses information collected by the in-store test set and data entered by the customer into the system, information from reference user databases and existing sleep set databases, to recommend sleep sets that have the best probability of providing the best sleep quality for that person. Detailed description:

[0064] A) The in-store trial set:

[0065] The intelligent sleep system according to the present invention comprises the mattress, which in turn comprises

[0066] i. A control box with one or more pressure sensors,

[0067] i. A BCG sensor (ballistocardiography sensor) connected to the control box, iii. Air chambers connected to the control box,

[0068] with the pressure sensor(s) configured to measure the pressure in the air chambers.

[0069] According to a particular embodiment, the control box comprises;

[0070] - an air pump, configured to inflate the air chambers; and

[0071] - solenoid valves, configured to control the opening or closing of the air chambers, preferably one solenoid valve per air chamber.

[0072] According to one specific embodiment, the control box includes a single pressure sensor. In this case, the solenoid valves are opened and closed to measure the pressure in each chamber individually. According to an alternative embodiment, the control box includes one pressure sensor for each air chamber, allowing the pressure in all air chambers to be measured simultaneously.

[0073] According to one particular embodiment, the prescription system mattress includes two beds. In this case, the mattress can have a set of air chambers and a BCG sensor for each bed, so that the prescription can be administered to two people simultaneously.

[0074] The BCG sensor (typically an accelerometer / gyroscope) measures, collects, and transmits vital signs and physical data such as heart rate (HR), respiratory rate (RR), heart rate variability (HRV) (which correlates with stress levels), relative stroke volume (how much blood the heart pumps), and movement (whether the bed is empty or the patient is in bed, and whether the patient is moving). The same sensor also estimates a person's fitness level by measuring their heart rate, respiratory rate, and heart rate variability. It can also detect muscle tension. By taking measurements while the air chambers are inflating or deflating, it assists in determining the optimal mattress or pillow settings to improve the user's comfort or respiratory health.It can also support the baseline prescription process by linking measured parameters with certain inclinations that help improve respiratory (such as: apnea, snoring) or cardiovascular health.

[0075] The pressure variation caused by the user in the main usual resting positions is determined, in this particular implementation in supine and lateral positions:

[0076] (i) The user initially positions themselves supine on the test unit in the store. The mattress may integrate, for example, six air chambers grouped into 4 zones (shoulders, chest, hips and thighs), and at least one pressure sensor. The number of air chambers may be greater, depending on the needs.

[0077] (i) For each position (supine and lateral), the system performs two consecutive measurements in each chamber:

[0078] - Baseline measurement: pressure recorded with the user at rest.

[0079] - Measurement after inflation: pressure recorded after simultaneously inflating all chambers for a fixed interval (e.g., 1 or 2 seconds), followed by a short stabilization time.

[0080] (iii) The difference between the two values ​​reflects how the user's body interacts with the mattress in that position.

[0081] These pressure variations allow us to characterize the response of the mattress-user system and obtain relevant information about the distribution of loads in the main usual resting positions.

[0082] The curvature of the user's spine is determined by the following steps: i. When the customer is positioned supine in the in-store fitting, the air chambers are inflated sequentially along the user's spine for a specific time. Preferably, the air chambers are inflated for 8 to 15 seconds, and more preferably for 10 seconds, depending on the inflation equipment used and its power or the size of the chambers;

[0083] i. The pressure reached in each chamber is measured, which depends on the resistance exerted by each area of ​​the person's body.

[0084] iii. The pressure values ​​achieved are transformed into shape values ​​according to a correspondence between pressure and shape. The shape values ​​form a curve that represents the curvature of the back.

[0085] Therefore, by knowing the pressures reached in the air chambers, the shape of the person's back can be mimicked (Fig. 2). This data is used to recommend the mattress by zones.

[0086] The air chambers are embedded transversely in the mattress, in the support layer of the test assembly. As is known in the art, in common embodiments a mattress comprises at least a comfort layer and a support layer, arranged in a superimposed manner.

[0087] The comfort layer provides the mattress's comfort setting and is designed to allow local adaptation to the user's body surface. Meanwhile, the support layer establishes the mattress's structural support profile and is intended to provide the necessary load-bearing capacity to support the user's body weight.

[0088] The distinction between the materials used in comfort and support layers is well-known to any expert in the mattress industry, as both layers serve different purposes and are therefore composed of materials with different mechanical properties. While the support layer aims for greater resilience and mechanical stability to support the body's weight and limit overall mattress deflection, the comfort layer primarily serves to adjust the feeling of comfort and the local distribution of pressure.

[0089] Unlike systems where air chambers are primarily located in the comfort layer, this configuration prioritizes measurements related to the mattress's structural support. This allows for a more representative characterization of the interaction between the user's load and the overall mechanical response of the sleep system. Placing the air chambers in the support layer reduces the measurements' dependence on the local properties of the comfort layer, which in current commercial mattresses is typically highly optimized for surface adaptability using viscoelastic or similar materials.

[0090] Additionally, the system may include a mechanism configured to reduce the dependence of measurements on the comfort layer, compensating for variations due to the type, condition, or degree of deterioration of the test mattress padding. In this way, the system provides self-calibration, allowing for more stable and comparable results across different test mattresses, measurement sessions, or locations.

[0091] Preferably, according to one particular embodiment, the air chambers are tube-shaped. More preferably, the tubes have a rectangular cross-section. According to other particular embodiments, the tubes may have other types of cross-sections, for example, circular, oval, etc.

[0092] The system also includes the user interface for customer data entry, for example, anthropometric measurements such as gender, height, weight and body circumferences, age, temperature sensitivity, known health or sleep problems, sleep habits, or sleeping posture preferences, in addition to personal data such as name, surname, email or telephone.

[0093] Preferably, the intelligent sleep system according to the present invention also includes a pillow. The pillow uses the same process as the mattress. This pillow comprises one or more air chambers, preferably two, connected to a control box with one or more pressure sensors configured to measure the pressure in the air chambers. The air chamber(s) are inflated with air, the pressure reached in each air chamber is measured, and a morphological curve is defined in the same way as in the case of the mattress, to represent the cervical shape. This data, along with the user's preferred sleeping position and other factors such as whether or not they have respiratory conditions, etc., are used to recommend the type of pillow.Additionally, the use of the ballistocardiography (BCG) sensor, which takes measurements while these air chambers are inflating or deflating, could determine which pillow configuration is optimal for improving the user's respiratory health, combining that information with all the parameters listed above in order to make a better recommendation.

[0094] Preferably, the intelligent sleep system according to the present invention also comprises an adjustable base (e.g., an adjustable, articulated, or motorized base). The base moves during the process but does not take measurements. However, this is important because the system can also recommend bases based on the user's needs, according to data entered through the user interface. Different bases are recommended, for example, based on the user's preferred sleeping position, such as sleeping on their back or side, or based on medical conditions, such as reflux, migraines, or sleep apnea.Additionally, the base can support the prescription process by studying, through parameters measured by the ballistocardiography (BCG) sensor, how certain inclinations help improve respiratory or cardiovascular health, adding those values ​​when prescribing the base (recommending, for example, one that has anti-Trendelenburg movement if the user has stated that they sleep on their side and have shown significant improvements in their vital signs during the prescription process with the base tilted to that position).

[0095] Optionally, the system includes a temperature sensor to measure body temperature and, in combination with the information collected by the user, prescribe mattresses with active or passive thermoregulating elements, or a specific set of sheets or a duvet. B) The system for classifying mattresses and pillows (prescribable sleep set data):

[0096] This is a methodology or database for classifying commercial sleep sets (bases, mattresses and pillows) according to their properties and physical characteristics (a series of standards), mainly the level of comfort, but also thermal conductivity, number of zones, shape (in the case of a pillow), number of movements and characteristics (in the case of a base), etc.

[0097] Any known classification methodology can be used; for example, there are ISO or similar standards to determine the level of comfort and thermal conductivity. For pillows, there are several common types such as regular, cervical, and different heights, and for bases, the movements / inclinations they can perform are known.

[0098] To characterize a mattress, since mattresses can come from different manufacturers who might use different characterization criteria, a machine can be used that can be adjusted to measurements according to various international standards (e.g., ISO 23769). Pillows are characterized using similar standards (e.g., ISO 2439). The machine can be located in the retailer's warehouse, at their factory (if they are also the manufacturer), or at a third-party facility, to ensure uniform characterization of all necessary components.

[0099] The Hs parameter, defined in ISO 23769, represents the objective firmness of a mattress, obtained through mechanical tests that measure the relationship between applied force and deflection produced in the entire sleep system. This value integrates the combined behavior of all mattress layers, including both comfort materials (top foams, padding, etc.) and support materials (springs, high-density structural foams, and other cores). The result is a standardized index on a scale of 1 to 10, reflecting the overall response of the mattress under load.

[0100] While the Hs value is an objective parameter, the user's perception of firmness depends on individual biomechanical factors, primarily body weight and habitual resting position. In general:

[0101] Users of lower weight perceive the mattress as firmer, so their optimal Hs range shifts towards lower values.

[0102] Heavier users generate greater deflection and perceive the mattress as softer, requiring higher Hs values.

[0103] The sleeping position modifies the level of adaptability required:

[0104] - Sleeping on your side requires less firmness (greater immersion capacity).

[0105] - Sleeping on your back requires medium firmness. - Sleeping on your stomach requires greater firmness to avoid lumbar hyperextension.

[0106] Based on these principles, it is possible to establish a predictive model that determines the optimal firmness level for a specific user, based on their anthropometric and postural characteristics. This model allows for personalized mattress selection or the adjustment of variable firmness systems, ensuring proper spinal alignment and balanced pressure distribution.

[0107] C) The database with historical data of real people

[0108] The intelligent sleep set prescription system according to the present invention comprises a set of reference data corresponding to respective reference users, comprising one or more of the following types of data: i. Sleep quality metrics,

[0109] i. Reference user profiles,

[0110] iii. Health data,

[0111] iv. Sensitivity to temperature,

[0112] v. Characteristics of the rest set used.

[0113] This is a database containing historical data from real people. It stores users' sleep quality metrics (quantified sleep quality), user profiles (including physical characteristics, weight distribution in main sleeping positions, back shape, health status (including any illnesses), and temperature sensitivity (temperature preferences), along with the characteristics of the sleep sets used by these individuals, including their respective Hs and Hs_final parameter values. In short, it contains all the relevant information needed to determine the correct choice of sleep set.

[0114] The intelligent sleep set prescription system according to the present invention also comprises the data of prescribable sleep sets, as defined above in point B), and the programmable processing means, in order to make the final prescription of the most optimal set according to the characteristics and needs of the particular customer.

[0115] According to a second aspect of the present invention, represented in Fig. 1, a computer-implemented prescription procedure for a rest set is described, comprising the following actions:

[0116] a. User data is entered through a user interface selected from the group comprising one or more of the following data types:

[0117] i. Anthropometric data,

[0118] i. Sleep habit data,

[0119] iii. Data related to health problems,

[0120] iv. Sensitivity to temperature;

[0121] b. The user lies down on a mattress comprising a control box with one or more pressure sensors, a BCG sensor (ballistocardiography sensor) connected to the control box, and air chambers connected to the control box, the pressure sensor(s) being configured to measure the pressure in the air chambers;

[0122] c. The pressure variation caused by the user is determined using the following steps:

[0123] i. The user is positioned supine;

[0124] i. A baseline measurement is taken in each chamber with the user at rest; iii. The air chambers are inflated simultaneously for a set time, preferably for 1 or 2 seconds;

[0125] iv. Another pressure measurement is taken in each chamber;

[0126] v. Steps (ii)-(iv) are repeated with the user in a lateral position;

[0127] vi. The pressure increase for each chamber is calculated for both postures; d. a session load (S) is calculated as an average value of the pressure increases measured in the air chambers in the supine and lateral postures;

[0128] e. The BCG sensor measures the user's vital signs;

[0129] f. The curvature of the user's back is determined using the following steps:

[0130] i. The user is positioned supine;

[0131] i. The air chambers are inflated sequentially along the user's back for a specific time. Preferably, each air chamber is inflated for the same amount of time as the others, from 8 to 15 seconds, and more preferably 10 seconds;

[0132] iii. The pressure reached in each chamber is measured, which depends on the resistance exerted by each area of ​​the person's body;

[0133] iv. The pressure values ​​achieved are transformed into shape values ​​according to a correspondence between pressure and shape. The shape values ​​form a curve that represents the curvature of the back.

[0134] According to a particular embodiment, the correspondence between pressure and shape comes from a database, obtained, for example, from previous tests. The following explains how the pressure variation caused by the user is determined according to step c) of the procedure:

[0135] The user initially positions themselves supine on the mattress, which according to a preferred embodiment integrates six air chambers grouped into 4 zones (shoulders, thorax, hips and thighs), and at least one pressure sensor.

[0136] First, a baseline pressure measurement is taken in each chamber with the user at rest. Then, the air chambers are inflated simultaneously for a predetermined time, preferably 1 or 2 seconds (corresponding, for example, to 2 or 3 mbar). After a brief stabilization period, another pressure measurement is taken in each chamber. The user is then instructed to change position to a side-lying position. Again, a baseline pressure measurement is taken in each chamber with the user at rest. Then, the air chambers are inflated simultaneously for a predetermined time, preferably 1 or 2 seconds (corresponding, for example, to 2 or 3 mbar). After a brief stabilization period, another pressure measurement is taken in each chamber. The pressure increase for each chamber is calculated for both positions.

[0137] These measurements allow characterizing the pressure variation caused by the user in the main usual resting positions, according to the present invention in supine and lateral positions.

[0138] The following explains how the curve representing the curvature of the back is formed from the shape values, according to a particular realization:

[0139] - The shape values ​​are graphically transferred onto each corresponding camera, thus obtaining graphic shape points.

[0140] - The curvature of the user's back corresponds to a curve formed by the graphic shape points.

[0141] The concept of resistance mentioned in step f) iii is explained in more detail below. The resistance exerted by each area of ​​a person's body during the inflation of the air chambers depends on the shape of the back in relation to the corresponding area of ​​the mattress. For example, depending on the mattress of a particular prescription system: If the resistance is zero (because no one is lying on it), it can reach a pressure of 180 mbar in 15 seconds of inflation.If someone is lying down, in each air chamber, after 15 seconds of inflation, a pressure is reached that depends on the curve of the back; for example, a person with a pronounced dorsal curvature will exert more resistance (and, consequently, more pressure will be produced in the air chamber or chambers on which they rest), than one with a straighter upper back; a person with a more pronounced lumbar curvature will generally exert less resistance in that area (and, consequently, less pressure will be produced in the air chamber or chambers on which they rest) than another person with a less pronounced lumbar curvature; etc.

[0142] The sequential inflation procedure of the air chambers characterizes the interaction between the user and the mattress of the test set and improves the reliability of the firmness calculation.

[0143] Furthermore, sequential inflation allows for the calculation of the so-called PSI (Padding Softness Index), a global indicator of the mechanical response of the test mattress's padding. The use of PSI provides continuous self-calibration of the system, reducing dependence on the type or condition of the test mattress, padding tolerances (for example, ISO standards accept a firmness variability of ±15% between batches of the same foam quality), and ensuring consistent measurements between different stores and testing sessions.

[0144] g. Using an AI proximity algorithm, preferably a k nearest neighbors (KNN) algorithm or similar, the data obtained in the previous stages a), c), d), e) and f) are compared with a set of reference data corresponding to respective reference users, said reference data comprising one or more of the following types of reference user data:

[0145] i. Sleep quality metrics,

[0146] i. Reference user profiles,

[0147] iii. Health data,

[0148] iv. Sensitivity to temperature,

[0149] v. Characteristics of the rest set used,

[0150] and the user is assigned the nearest reference user;

[0151] h. The rest set is prescribed to the user by assigning the rest set used by the nearest reference user.

[0152] It is important to note that the database with historical data from real people (reference users) provides data that actually determines the recommendation. In the procedure of the present invention, first, the user's data is entered through a user interface, the measurement is taken when the user lies down on the mattress, and the AI ​​algorithm (which may be a KNN or similar https: / / www.ibm.com / es-es / topics / knn) searches the historical database to determine which of the previously classified mattresses, bases, and pillows is likely to produce the best results in sleep quality. D) The software based on the AI ​​algorithm, preferably the k-nearest neighbors (KNN) algorithm or similar:

[0153] The k-nearest neighbors (KNN) algorithm is a nonparametric supervised learning classifier that uses proximity to make classifications or predictions about the grouping of an individual data point. It is one of the most popular and simplest classification and regression classifiers used in machine learning today.

[0154] In the case of the present invention, the reference user database is a database that has been created over time based on different sensors sold to multiple international manufacturers. In this way, the following data has been stored in the cloud:

[0155] - the reference user data (their physical characteristics, gender, age, sleeping position, health problems / data, temperature preferences, etc.), i.e., the same data that the user enters through a user interface of the system of the present invention,

[0156] - the data determined by the in-store test set (BCG sensor measurements, pressure variation caused by the user in the main usual resting postures (load distribution in the main resting postures), calculated Hs_final, curvature / shape of the back),

[0157] - What mattress does each reference user have with their respective Hs value, and

[0158] - What is the sleep quality of each reference user (quantified sleep quality)?

[0159] In this way, a portfolio of products (mattresses, bases, and pillows) is now available, which were previously characterized and recommended based on input from an AI algorithm. These tests allowed for the accumulation of several million hours of sleep data, and the configuration of a specific sleep set was correlated with the user's sleep quality. Furthermore, this data provided the following information: the customer's posture, weight, height, age, physical conditions, country, etc. With all this data, it is possible to determine which sleep set is best for a given individual.

[0160] Thus, with this reference user data, the k nearest neighbors (KNN) algorithm can correlate data on a reference user's physical characteristics, the sleep set they use, and the resulting sleep quality they have, to find the optimal sleep set from among those existing for a new user based on the nearest neighbor.

[0161] In other words, the algorithm processes the information collected by the test set in the store and the data that the customer enters into the system, the information from the reference user databases and the existing sleep set databases to recommend the sleep sets that have the best probability of providing the best sleep quality to that person (customer or new user).

[0162] Algorithms are used (e.g., the k nearest neighbors (KNN) algorithm) to process the test set data and historical data to recommend the best sleep set from those previously ranked, which will provide the best sleep quality for a new user.

[0163] Calculation of the Hs value for a user (final Hs):

[0164] a) Calculation of a session load (S)

[0165] Based on pressure variations relative to baseline pressure, the pressure increase for each chamber is calculated for both positions. The session load (S) corresponds to an average value of the pressure increases measured for each chamber (a single aggregated value) and provides a functional estimate of the relative subsidence induced by the user, without requiring anthropometric data such as weight or height.

[0166] b) Session reference calculation (sRef)

[0167] To contextualize the session load value (S) in relation to other previously measured users, a value called the session reference (sRef) is obtained.

[0168] This value can be calculated as the local median session load (S) obtained from previous users with measurements close to the current session. If there are not enough neighboring sessions—for example, fewer than 3 or 4—the system uses a market seed value, previously calculated from measurements taken from a representative set of users that reflect the most common ergotypes in the corresponding market or geographic region.

[0169] c) Transformation relative to preliminary firmness

[0170] In this step, a first approximation of the Hs value is made, called preliminary firmness (Hs_raw). A dimensionless relative value is calculated:

[0171] r = S / sRef, where:

[0172] S is the session load in the current session calculated in step a)

[0173] sRef is the reference value from previous sessions or the market seed value, calculated in step b)

[0174] Next, a nonlinear mapping is applied to convert this relative value r into Hs_raw on a controlled scale (e.g., [1,10] according to ISO 23769). This conversion curve is known and has been previously developed through experience, data collected from existing users, sleep quality data collected on different mattresses with different Hs values, etc.

[0175] The result is an Hs_raw value based exclusively on the user's mechanical interaction with the test mattress.

[0176] d) Calculation of final firmness (Hs_final)

[0177] Starting from Hs_raw, the system makes additional adjustments to adapt the final firmness to the user's body profile, posture, and possible therapeutic needs.

[0178] Adjustment for body morphology:

[0179] The morphology is obtained by analyzing the pressure differences detected during sequential inflation. An anatomical "prominence" is detected when the relative difference between zones exceeds a configurable threshold (typically 20%, based on ergonomics and orthopedics studies).

[0180] Examples:

[0181] - Prominent shoulders + sideways posture — ► reduced firmness (more padding) to allow for greater immersion.

[0182] - Prominent hips + supine posture > controlled adjustment to avoid excessive pelvic tilt.

[0183] Posture adjustment:

[0184] The relative weight of the supine / lateral measurements is adjusted according to the declared preferred position:

[0185] - Side sleeper — ► greater weight to the side measurement.

[0186] - Supine sleeper — ► greater weight in the supine measurement.

[0187] Adjustment for clinical problems:

[0188] If the user reports shoulder, hip, or lower back pain, specific corrections are applied that soften or strengthen the firmness locally.

[0189] The final firmness Hs_final is limited on a standard scale ([1-10] according to ISO 23769), representing the optimal firmness for that user according to their biomechanical interaction, posture and clinical conditions.

[0190] Other settings:

[0191] In addition to the steps described above, several adjustments are made during the prescribing process to strengthen the system and ensure its repeatability within the user, whether at the same point of sale or at distant points of sale. These adjustments reduce the system's dependence on the materials used in the comfort layer of the test mattress, adjust the prescription ranges to the preferences and / or availability of mattresses in a given market, allow the prescribing system to learn from use and adapt its prescriptions to the different body shapes of users in a specific geographic area, and enable better recommendations for individuals with irregular or problematic body shapes. All these adjustments resolve the problems present in current models, which are a frequent source of complaints from points of sale using prescribing systems.

[0192] Some examples of these adjustments:

[0193] Session load adjustment (S) to compensate for intra-user deviations due to different positioning on the test mattress:

[0194] When a user performs several consecutive measurements (for example, a user who takes a test one day and returns another day to confirm the purchase), the system can reduce variability between repetitions—due to different positioning, postural micro-adjustments, or initial pressure variations—using techniques based on median absolute deviation (MAD), thus obtaining a more stable session load (S) value. Adjustment of the session load (S) level by the padding level and its deterioration:

[0195] During inflation, each air chamber inflates for a fixed time while the user lies down. During this interval, the increase in pressure relative to the initial state is measured. This increase is directly related to the mattress padding's ability to absorb pressure before transferring it to the chambers.

[0196] Based on these increases, the Padding Softness Index (PSI) is defined as the average of the pressure increases in all chambers:

[0197] - Low PSI — ► very soft mattress (absorbs some of the pressure).

[0198] - High PSI — ► rigid mattress or one with thin padding (transmits pressure more directly).

[0199] This PSI constitutes a global indicator of the mechanical response of the mattress padding of the test set in store.

[0200] Once the baseline value of the session load (S) has been calculated, as the average of the pressure increases in supine and lateral positions relative to the baseline measurement, the PSI is used to normalize this value.

[0201] This correction is made proportionally to the ratio between the measured PSI and the reference PSI (for example, that corresponding to the mattress used as a standard in the factory or laboratory):

[0202] - Softer mattress (lower PSI than the reference) - the S value increases proportionally, as some of the pressure is dissipated in the padding.

[0203] - Stiffer mattress (higher PSI) - the S value is reduced, because the load is transmitted more directly.

[0204] The use of PSI provides continuous self-calibration of the system, reducing dependence on the type or condition of the mattress in the test set in store, on the tolerances of the padding (for example, in foam, a firmness variability of +-15% between batches of the same foam quality is accepted by ISO standard) and ensuring consistent measurements between different stores and sessions.

[0205] Updating the seed value for calculating Sref:

[0206] Optionally, the system allows for the periodic recalculation of this seed value as new sessions are registered, so that it automatically adapts to the actual evolution of the market or user flow. This is especially relevant whenever the system is implemented in a new market with ergotypes different from those with which the system was trained. For example, if the system is launched with the seed value based on European users, it is likely that ergotypes from the US or Asia will require this seed value to be updated in order to normalize the session load (S) values ​​to the specificities of those geographic areas.

[0207] Curve mapping adjustment that transforms r into Hs_raw:

[0208] This curve is modulated by three main variables:

[0209] - Dead zone: range in which small variations in r do not alter the value of Hs, to avoid excessive sensitivity to noise or micro-adjustments of posture.

[0210] - Exponent (y): controls the curvature of the transformation and the sensitivity levels to user deviations from the reference sRef.

[0211] - Gain (A): adjusts the final amplitude of Hs_raw to suit the available range of mattresses on the market (for example, homogeneous markets where all mattresses are of medium-high firmness such as Spain or Asia versus more diverse markets such as the USA or Australia).

[0212] Once Hs_final has been calculated, the system moves on to the personalized prescription phase of the complete rest set.

[0213] Mattress selection

[0214] The mattresses in the database are filtered out if their firmness value matches or is closest to the calculated Hs_final.

[0215] Among these, one or more models are recommended, taking into account certain variables such as:

[0216] - optimized for declared clinical problems or thermal sensitivity,

[0217] - usual sleeping position,

[0218] - thermal sensitivity,

[0219] - include differentiated zones according to detected morphology (for example, a softer shoulder area),

[0220] - have shown better performance in users with similar profiles (based on satisfaction history or sleep quality).

[0221] Pillow selection

[0222] Optionally, the system incorporates a pillow with sensors that allow for pressure measurements in the head and neck area. These measurements, along with parameters such as the declared sleeping position, the load calculated by the mattress in the shoulder area, declared health or thermoregulation issues, and the recommended final head circumference (Hs_final), are used to determine the ideal pillow height and firmness. For example, it is known that a firmer mattress generally requires a higher or firmer pillow, a softer mattress a lower and / or softer pillow, and someone who sleeps on their back or stomach will need a lower pillow than someone who sleeps on their side.

[0223] Furthermore, if the pillow features active elements such as air chambers that allow its shape to be modified, the use of the ballistocardiography (BCG) sensor, which takes measurements while these air chambers are inflating or deflating, could determine which pillow configuration is optimal for improving the user's respiratory health, combining this information with all the parameters listed above in order to make a better recommendation.

[0224] This pillow measurement system could be replaced by any other known system for recommending pillows, for example an optical device (video camera) capable of measuring shoulder width, head shape, etc., and which is integrated into the recommendation process or is done independently.

[0225] Base selection

[0226] Optionally, depending on the measurements taken and the problems reported, a fixed or adjustable base is recommended, for example electrically, e.g. in situations of lower back pain, snoring or reduced mobility.

[0227] The base can be integrated into the prescription process, similarly to what was described in the case of the pillow, by studying, through the parameters measured by the ballistocardiography (BCG) sensor, how certain inclinations of the base help to improve respiratory or cardiovascular health, adding those values ​​when prescribing the base (recommending, for example, one that has anti-Trendelenburg movement if the user has declared sleeping on their side and has shown significant improvements in their vital signs during the prescription process with the inclination of the base to that position).

[0228] In-store testing process:

[0229] Step 1 (represented in the “MAPPING SYSTEM” box in Fig. 1):

[0230] The customer lies supine on the smart mattress. Before, during, or after the customer enters the data (gender, preferred position, etc.). The BCG sensor begins measuring, detecting presence to avoid false readings, and assesses fitness level through heart rate (HR), respiratory rate (RR), and heart rate variability (HRV), collecting objective physical data.

[0231] Step 2 (represented in the “MAPPING SYSTEM” box in Fig. 1):

[0232] The air chambers are inflated simultaneously for a set time (typically 1 or 2 seconds). A baseline measurement is taken in each chamber with the user at rest, and once inflated, another pressure measurement is taken in each chamber. The pressure increase for each chamber is calculated, and the procedure is then repeated with the user lying on their side. These measurements allow for the characterization of the load level generated by the user in the main common resting positions (supine and lateral).

[0233] Step 3 (represented in the “MAPPING SYSTEM” box in Fig. 1):

[0234] The air chambers, which are initially empty, are inflated sequentially for a certain time (8 to 15 seconds, preferably for 10 seconds), so that all the air chambers are inflated. Each one is inflated for a certain time depending on the resistance that each area of ​​the person's body exerts (action-reaction principle). The pressure level reached in each one is measured and the result is stored; so that by knowing the pressures reached, the shape of the person's back can be estimated / mimicked as described above, key information for making recommendations for zoned mattresses (areas that require additional support).

[0235] Step 4:

[0236] Mattress recommendations: These are made using a recommendation algorithm, which suggests a zoned mattress if it believes that will provide the best results. Zoned mattresses essentially offer different levels of firmness depending on the individual's body shape or specific needs.

[0237] Step 5:

[0238] Pillow recommendation: a process similar to steps 2 and 3 for the mattress, mapping the cervical shape, additional information to prescribe the appropriate pillow.

[0239] Step 6: Base recommendation: For example, if the person has no condition, the recommendation could be a fixed base; if the person has respiratory, circulatory, gastric conditions, etc., and sleeps on their side, it could be a base with anti-Trendelenburg; if they have apneas or snoring, it could be a base with head tilt.

[0240] Step 7:

[0241] Accessory recommendations: For example, recommendations could be made about duvets (e.g., for people who get cold easily), sheets (e.g., sheets with thermally conductive elements such as PCM), mattress toppers (e.g., mattress toppers with active cooling elements), humidifiers (e.g., for people with respiratory problems), etc.

[0242] So all the information is taken into consideration, in combination with the information entered by the client, so that the AI ​​algorithm prescribes a particular sleep pattern that benefits certain health conditions. For example, it is known in the field that respiratory problems can improve when sleeping on one's side, or digestive problems when sleeping with one's back in the reverse Trendelenburg position.

[0243] In addition, the system can prescribe other elements, for example, depending on the user's conditions that are entered into the system through the user interface, such as pressotherapy incorporated or not in the mattress, vibration therapy or any other system that promotes relaxation to alleviate pain or reduce sleep latency.

Claims

AMENDED CLAIMS Received by the International Office on June 17, 2026 (17.06.2026) 1. Intelligent prescription system for a rest set comprising: a. a mattress comprising: i. a control box with one or more pressure sensors, i. a BCG sensor connected to the control box, iii. air chambers connected to the control box, with the pressure sensor or sensors configured to measure the pressure in the air chambers; b. a user interface configured to input user data, selected from the group comprising one or more of the following data types: - anthropometric data, - sleep habits data, - data related to health problems, - temperature sensitivity; c. a set of reference data corresponding to respective reference users, comprising one or more of the following data types: i. sleep quality metrics, i. reference user profiles, iii. health data, iv. sensitivity to temperature, v. characteristics of the rest set used; d. data on prescriptive rest sets; and e. a programmable processing medium configured to: i. receive user data entered through the user interface; i. Determine the pressure variation caused by the user by: - obtaining a baseline measurement in each chamber with the user at rest in a supine position, - the simultaneous inflation of the air chambers for a certain time, - obtaining another pressure measurement in each chamber, - repeating the above operations with the user in a lateral position, and - the calculation of a pressure increase for each chamber for both positions; iii. calculate a session load (S) as an average value of the pressure increments measured in the air chambers in the supine and lateral positions; iv. receive the user's vital signs measured by the BCG sensor; v. Determine the curvature of the user's back by: - the sequential inflation of the air chambers for a certain time in a longitudinal direction along the user's back, with the user in a supine position, - the measurement of the pressure reached in each chamber, which depends on the resistance exerted by each area of ​​the person's body, and - the transformation of the pressure values ​​achieved into shape values ​​according to a correspondence relationship between pressure and shape, forming the shape values ​​into a curve that represents the curvature of the back; vi. compare, using an AI proximity algorithm, such as the k nearest neighbors (KNN) algorithm or similar, the data obtained in operations i), iii), iii), iv) and v) with the reference data set corresponding to respective reference users; vii. assign the user a nearest reference user; and viii. Prescribe the rest set to the user by assigning the rest set used by the nearest reference user.

2. System according to claim 1, comprising a pillow with one or more air chambers connected to the control box.

3. System according to any of the preceding claims, wherein the air chambers are embedded transversely in the support layer of the mattress.

4. System according to any of the preceding claims, comprising an adjustable base.

5. System according to any of the preceding claims, comprising a temperature sensor.

6. System according to any of the preceding claims, wherein the air chambers are tube-shaped.

7. System according to any of the preceding claims, wherein the mattress includes two beds, each with a set of air chambers and with a BCG sensor for each bed.

8. Computer-implemented rest set prescription procedure comprising the following actions: a. User data is entered through a user interface selected from the group comprising one or more of the following data types: i. anthropometric data, i. sleep habit data, iii. data related to health problems, iv. sensitivity to temperature; b. The user lies down on a mattress comprising a control box with one or more pressure sensors, a BCG sensor connected to the control box, and air chambers connected to the control box, the pressure sensor(s) being configured to measure the pressure in the air chambers; c. The pressure variation caused by the user is determined by the following steps: i. the user is positioned supine; i. a baseline measurement is taken in each chamber with the user at rest; iii. the air chambers are inflated simultaneously for a predetermined time; iv. another pressure measurement is taken in each chamber; v. steps (ii)-(iv) are repeated with the user in a lateral position; vi. A pressure increase is calculated for each chamber for both positions; d. a session load (S) is calculated as an average value of the pressure increments measured in the air chambers in the supine and lateral positions; e. the BCG sensor measures the user's vital signs; f. The curvature of the user's back is determined using the following steps: i. the user is positioned supine; i. The air chambers are inflated for a certain time, sequentially in a longitudinal direction along the user's back; iii. The pressure reached in each chamber is measured, which depends on the resistance exerted by each area of ​​the person's body; iv. the pressure values ​​achieved are transformed into shape values ​​according to a correspondence relationship between pressure and shape, the shape values ​​forming a curve that represents the curvature of the back; g. using an AI proximity algorithm, such as the k nearest neighbors (KNN) algorithm or similar, the data obtained in the previous stages a), c), d), e) and f) are compared with a set of reference data corresponding to respective reference users, said reference data comprising one or more of the following types of reference user data: i. sleep quality metrics, i. reference user profiles, iii. health data, iv. sensitivity to temperature, v. characteristics of the rest set used, and the user is assigned the nearest reference user; h. The rest set is prescribed to the user by assigning the rest set used by the nearest reference user.

9. Method according to claim 8, wherein the user's vital signs measured by the BCG sensor are: heart rate, respiratory rate, heart rate variability, relative stroke volume and / or movement.

10. Method according to claim 9, wherein the BCG sensor calculates the fitness level.

11. Method according to any of claims 8 to 10, wherein at least one mattress, at least one pillow, or at least one base, or any combination thereof, is prescribed.

12. A method according to any one of claims 8 to 11, further comprising: - calculate a session reference value (sRef) associated with previously measured user session loads; and - determine a dimensionless relative value (r) as the ratio between the session load (S) and the session reference value (sRef).

13. Method according to claim 12, wherein the session reference value (sRef) is calculated as a local median of session loads (S) obtained from previous users with measurements close to the current session, and in the absence of sufficient neighboring sessions, a market seed value calculated from measurements made to a representative set of users is used.

14. A method according to any of claims 12 or 13, wherein a preliminary firmness (Hs_raw) is calculated from the relative value (r) by means of a non-linear transformation having: - a dead zone where variations in r do not alter the value of Hs_raw; - an exponent (y) that controls the curvature of the transformation; and - a gain (X) that adjusts the final amplitude of Hs_raw to a predetermined firmness range.

15. A method according to any of claims 12 to 14, further comprising: - calculate a cushioning index (PSI) as the average of the pressure increases in all air chambers during inflation with the user lying down; and - Correct the session load (S) based on the ratio between the padding index (PSI) and a reference padding index (PSI_ref), increasing S when PSI is less than PSI_ref and reducing S when PSI is greater than PSI_ref.

16. A method according to any of claims 12 to 15, wherein the seed value used for calculating the session reference value (sRef) is periodically updated as new sessions are recorded, adapting to the actual evolution of the market or the flow of users in a given geographical area.

17. A method according to any of claims 12 to 16, wherein a final firmness (Hs_final) is calculated from the preliminary firmness (Hs_raw) by: - ​​an adjustment based on body morphology obtained from the pressure distribution of the air chambers, detecting anatomical prominences when the relative differences between areas exceed a predetermined threshold; - an adjustment based on the user's declared preferred sleeping position, modifying the relative weight of the supine and lateral measurements; and - an adjustment based on clinical problems declared by the user relating to shoulders, hips or lumbar region, applying specific local firmness corrections.

18. A method according to any of claims 8 to 17, wherein the prescription of the sleep set comprises selecting at least one mattress from the database of prescribable sleep sets by filtering those mattresses whose firmness value matches or is closest to the final firmness value (Hs_final), and prioritizing the models that: - are optimized for declared clinical problems or thermal sensitivity; and / or - feature differentiated zones adapted to the detected morphology; and / or - have shown better performance in users with similar profiles.

19. Method according to any of claims 8 to 18, wherein the prescription further comprises the selection of a pillow, using for this purpose: - parameters derived from the cervical shape obtained by means of air chambers in the pillow or by means of an optical system; - the declared sleeping position; - the load calculated by the mattress in the shoulder area; and - the final firmness value (Hs_final) recommended for the mattress.

20. A method according to any of claims 8 to 19, wherein the prescription further comprises the selection of a fixed or adjustable base based on data entered by the user, measurements from the BCG sensor and declared pathologies, recommending bases with certain inclinations when at least one of the following conditions is detected or declared: respiratory problems, circulatory problems, gastric problems, snoring, sleep apnea, lower back pain or reduced mobility.

21. System according to any of claims 1 to 7, wherein the programmable processing medium is configured to: - calculate the session load (S), the session reference value (sRef), the padding index (PSI), and the final firmness (Hs_final) according to any of claims 12 to 17; and - Select, from the data of prescribable sleep sets, at least one mattress, optionally at least one pillow and, optionally, at least one base, based on the final firmness value (Hs_final), the curvature of the back and the user's health and sleep habits data.

Images