Intelligent sleep state analysis and dynamic optimization system based on multi-dimensional biological data

Abstract

The invention relates to the technical field of sleep quality, and discloses an intelligent sleep state analysis and dynamic optimization system based on multi-dimensional biological data, the system comprises a sleep monitoring module, a sleep analysis module, an environment optimization module and a data processing module, the sleep monitoring module collects biological data and environment data according to a sensing device, and the data processing module processes the biological data and the environment data. The sleep analysis module obtains a sleep state by adopting a threshold logic algorithm based on the target biological data and predicts a biological clock sequence of the current sleep state by adopting an LSTM model according to historical sleep state data, and the environment optimization module adjusts a light mode and a playing mode of the sleep equipment based on the environment data. The vibration mode of the sleep equipment is changed according to the time of the change trend, the data processing module judges whether the light mode of the sleep equipment is correct or not according to historical adjustment data, if not, the light mode of the sleep equipment is re-determined through the clustering algorithm, and the sleep environment is dynamically adjusted, and the sleep quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of sleep quality, and in particular to an intelligent sleep state analysis and dynamic optimization system based on multi-dimensional biological data. Background Art

[0002] With the accelerated pace of life in modern society, sleep problems have become a common health challenge worldwide. According to the World Health Organization, more than 30% of the population has varying degrees of sleep disorders, which not only affects individual health, but also puts pressure on social and economic development. Traditional sleep management methods usually rely on a single behavioral intervention, such as fixed sleep time, auxiliary drugs or environmental adjustments, etc., lacking real-time feedback support for individual biological data, making it difficult to achieve personalized sleep optimization. Although the sleep system in current smart home devices can provide basic monitoring functions, the actual utilization rate of these data is insufficient and can only be used as a static reference, failing to provide dynamic environmental adjustment or personalized intervention measures.

[0003] Therefore, how to provide an intelligent sleep state analysis and dynamic optimization system based on multi-dimensional biological data is a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0004] In view of this, the present invention proposes an intelligent sleep state analysis and dynamic optimization system based on multi-dimensional biological data, aiming to solve the problems of insufficient utilization of sleep data and lack of real-time dynamic adjustment and adaptability.

[0005] In one aspect, the present invention provides a dynamic optimization system based on multidimensional biological data, comprising:

[0006] Sleep monitoring module, sleep analysis module, environment optimization module and data processing module;

[0007] The sleep monitoring module is configured to be electrically connected to the sensing device, collect biological data and environmental data according to the sensing device, and pre-process the biological data to obtain target biological data;

[0008] The sleep analysis module is configured to use a threshold logic algorithm to obtain a sleep state based on the target biological data, use an LSTM model to predict a biological clock sequence of a current sleep state based on historical sleep state data, and obtain a change trend of the current sleep state based on the biological clock sequence;

[0009] The environment optimization module is configured to dynamically adjust the sleep device, the environment optimization module is electrically connected to the sleep device, and the environment optimization module includes a sleep aid unit and a wake-up unit, the sleep aid unit determines whether the current sleep state is within the relaxation index based on the target biological data or adjusts the light mode and the play mode of the sleep device based on the environment data according to the change trend of the current sleep state, the wake-up unit adjusts the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and changes the vibration mode of the sleep device according to the time of the change trend;

[0010] The data processing module is configured to determine whether the lighting mode of the sleep device is correct based on historical adjustment data. If it is incorrect, a clustering algorithm is used to re-determine the lighting mode of the sleep device, and a sleep optimization report is established based on the adjustment results of the sleep device, target biological data, and the changing trend of the current sleep state.

[0011] Furthermore, when the biological data and environmental data are collected according to the sensing device and the biological data is pre-processed to obtain the target biological data, it includes:

[0012] The sensing device includes: a wearable device and a sensor;

[0013] Collect the biological data through the wearable device, and collect the environmental data through the sensor;

[0014] The biological data includes heart rate data, respiratory rate data and body movement times;

[0015] The environmental data includes environmental humidity data and environmental light intensity;

[0016] The heart rate data and the respiratory rate data are preprocessed, wherein the preprocessing includes data cleaning and standardization, and a target heart rate and a target respiratory rate are obtained based on the results of the preprocessing.

[0017] Furthermore, when the sleep state is derived based on the target biological data using a threshold logic algorithm, it includes:

[0018] The target heart rate is recorded as HR, the target respiratory rate is recorded as RR, and the number of body movements is recorded as BM;

[0019] When BM>5 or HR>80, the current sleep state is judged as awake;

[0020] When 15≤RR≤20, and 60≤HR≤75, the current sleep state is determined to be the first sleep state;

[0021] When RR < 15 and HR < 60, the current sleep state is determined to be the second sleep state;

[0022] When the awake state, the first sleep state or the second sleep state is not satisfied, the current sleep state is determined as an eye movement state.

[0023] Furthermore, when the LSTM model is used to predict the biological clock sequence of the current sleep state based on the historical sleep state data, it includes:

[0024] Dividing the historical sleep state data into a sleep training set and a sleep test set;

[0025] Using cross validation combined with grid search to find model parameters of the LSTM model, establishing the LSTM model, fitting the LSTM model according to the sleep training set, and substituting the sleep test set into the LSTM model to calculate the pass rate of predicting the biological clock sequence;

[0026] When the pass rate reaches a preset pass rate threshold, the current sleep state is substituted into the LSTM model to predict the biological clock sequence of the current sleep state.

[0027] Furthermore, when the change trend of the current sleep state is obtained according to the biological clock sequence, it includes:

[0028] Extracting features of the biological clock sequence, the features including: the time of transition between different sleep states, the duration of the sleep state, and the change trend of the sleep state;

[0029] The change trend includes a primary trend and a secondary trend, the primary trend is that the first sleep state is converted into the second sleep state, and the secondary trend is that the second sleep state is converted into the first sleep state.

[0030] Further, when judging whether the current sleep state is in the relaxation index based on the target biological data or adjusting the light mode and the play mode of the sleep device based on the environmental data according to the change trend of the current sleep state, it includes:

[0031] The sleeping equipment includes: sleeping lights, sleeping speakers and humidifiers;

[0032] When HR≤75 and BM≤3, the sleep state is judged as being in the relaxation index, otherwise the sleep state is not in the relaxation index;

[0033] When it is determined that the sleep state is at the relaxation index or the change trend is a primary trend, the play mode of the sleep sound is adjusted to a white noise mode, the on / off state of the humidifier is adjusted according to the ambient humidity data, and the light mode of the sleep device is adjusted according to the ambient light intensity;

[0034] When the ambient humidity data is less than the set humidity threshold, the humidifier is turned on; when the ambient humidity data is greater than or equal to the set humidity threshold, the humidifier is not turned on;

[0035] When the ambient light intensity is greater than the set light threshold, the light mode of the sleep light is adjusted to the soft mode. When the ambient light intensity is less than the set light threshold, the light mode of the sleep light is adjusted to the fill light mode. When the ambient light intensity is equal to the set light threshold, the sleep light is not turned on.

[0036] Further, when adjusting the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and changing the vibration mode of the sleep device according to the time of the change trend, it includes:

[0037] The sleeping device also includes: a smart mattress;

[0038] When the change trend is a secondary trend, the light mode of the sleep light is adjusted to a bright mode, the play mode of the sleep audio is adjusted to a light music mode, and the vibration mode of the smart mattress is adjusted to a first vibration mode;

[0039] According to the time of conversion of the different sleep states, the exchange time of converting the second sleep state to the first sleep state is obtained, and when the data processing module determines that the second sleep state is not converted to the first sleep state within the exchange time, the play mode of the sleep sound is adjusted to the heavy metal mode, and the vibration mode of the smart mattress is adjusted to the second vibration mode.

[0040] Further, when judging whether the light mode of the sleep device is correct according to the historical adjustment data, it includes:

[0041] The historical adjustment data includes historical sleep state data of the same type as the current sleep state, historical change trends, and corresponding historical adjustment light mode data;

[0042] The standardized result of adjusting the light mode according to the change trend of the current sleep state and the sleep state relaxation index is compared with the historical light mode adjustment data to obtain a ratio Q;

[0043] When Q is equal to 1, it indicates that the light mode is correct, and it is determined that the light mode of the sleep light is not to be modified;

[0044] When Q is not equal to 1, it indicates that the lighting mode is incorrect, and it is determined that the lighting mode of the sleep light should be corrected.

[0045] Furthermore, if it is incorrect, a clustering algorithm is used to re-determine the light mode of the sleeping device.

[0046] Obtain representative data of each historically adjusted light mode data according to the historical adjustment data, and establish an aggregated data set by combining the representative data and the light mode of the current sleeping device;

[0047] Extracting feature vectors of each data in the aggregated data set and determining important features;

[0048] Determine the expected number of clusters k as 3 and initialize the parameters of the Gaussian distribution;

[0049] Calculating the probability that each data in the aggregated data set belongs to each Gaussian distribution to obtain a responsibility value;

[0050] The cluster with the largest responsibility value is selected as the re-determined lighting mode of the current sleeping device.

[0051] Compared with the prior art, the beneficial effects of the present invention are: by collecting and analyzing biological data and environmental data, dynamically adjusting the sleep equipment, creating a suitable sleeping environment, and improving the quality of sleep, the sleep equipment is dynamically adjusted according to each user's biological data, sleep status and historical sleep data, and the biological clock sequence is predicted by the LSTM model, which can accurately grasp the biological rhythm, so as to appropriately adjust the sleep equipment, continuously track the trend of changes in sleep status, and monitor and adjust the mode of the sleep equipment in real time, which not only improves the sleep quality, but also can correct the sleep equipment according to historical adjustment data, further improves the accuracy of sleep optimization, effectively reduces interference factors, thereby improving the overall health level, and generates detailed sleep optimization reports, improves the comprehensive utilization of data, and enhances the adaptability and flexibility of the system.

[0052] On the other hand, the present application also provides an intelligent sleep state analysis based on multi-dimensional biological data, which is used to apply the above-mentioned dynamic optimization system based on multi-dimensional biological data, including:

[0053] Collecting biological data and environmental data, and preprocessing the biological data to obtain target biological data;

[0054] Based on the target biological data, a threshold logic algorithm is used to obtain the sleep state, and based on the historical sleep state data, a LSTM model is used to predict the biological clock sequence of the current sleep state, and a change trend of the current sleep state is obtained based on the biological clock sequence;

[0055] Dynamically adjust the sleep device, determine whether the current sleep state is in the relaxation index based on the target biological data, or adjust the light mode and the play mode of the sleep device based on the environmental data according to the change trend of the current sleep state, adjust the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and change the vibration mode of the sleep device according to the time of the change trend;

[0056] Whether the lighting mode of the sleep device is correct is determined based on historical adjustment data. If incorrect, a clustering algorithm is used to re-determine the lighting mode of the sleep device. A sleep optimization report is established based on the adjustment results of the sleep device, target biological data, and the changing trend of the current sleep state.

[0057] It is understandable that the above-mentioned intelligent sleep state analysis and dynamic optimization system based on multi-dimensional biological data has the same beneficial effects, which will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0059] Figure 1 A schematic diagram of the structure of a dynamic optimization system based on multi-dimensional biological data provided by an embodiment of the present invention;

[0060] Figure 2 A schematic diagram of the structure of an environment optimization module provided by an embodiment of the present invention;

[0061] Figure 3 A schematic diagram of the structure of a sensing device provided in an embodiment of the present invention;

[0062] Figure 4 A schematic diagram of the structure of a sleeping device provided by an embodiment of the present invention;

[0063] Figure 5 A flowchart of intelligent sleep state analysis based on multi-dimensional biological data is provided in an embodiment of the invention. DETAILED DESCRIPTION

[0064] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0065] See also Figure 1-2 As shown, in some embodiments of the present application, a dynamic optimization system based on multidimensional biological data includes:

[0066] Sleep monitoring module, sleep analysis module, environment optimization module and data processing module;

[0067] The sleep monitoring module is configured to be electrically connected to the sensing device, collect biological data and environmental data according to the sensing device, and pre-process the biological data to obtain target biological data;

[0068] The sleep analysis module is configured to use a threshold logic algorithm based on target biological data to derive a sleep state, use an LSTM model to predict a biological clock sequence of a current sleep state based on historical sleep state data, and derive a change trend of the current sleep state based on the biological clock sequence;

[0069] The environment optimization module is configured to dynamically adjust the sleep device. The environment optimization module is electrically connected to the sleep device. The environment optimization module includes a sleep aid unit and a wake-up unit. The sleep aid unit determines whether the current sleep state is in a relaxation index based on the target biological data or adjusts the light mode and the play mode of the sleep device based on the environment data according to the change trend of the current sleep state. The wake-up unit adjusts the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and changes the vibration mode of the sleep device according to the time of the change trend.

[0070] The data processing module is configured to determine whether the lighting mode of the sleep device is correct based on historical adjustment data. If it is incorrect, a clustering algorithm is used to re-determine the lighting mode of the sleep device, and a sleep optimization report is established based on the adjustment results of the sleep device, target biological data, and the changing trend of the current sleep state.

[0071] Specifically, the sleep monitoring module is the basic module of the system, which is electrically connected to the sensing device to collect biological data and environmental data of the environment in real time. The sleep monitoring module effectively eliminates interference and errors between data by preprocessing biological data, improves the data quality and accuracy of target biological data, and provides reliable data support for subsequent sleep analysis and environmental optimization. The sleep analysis module is the core module of the system. It uses a threshold logic algorithm to judge the sleep state based on the target biological data, and uses historical sleep state data to train the LSTM (long short-term memory) model to obtain the biological clock sequence of the current sleep state. The LSTM model is a time series data prediction model based on deep learning. It can accurately predict the future sleep state and arrange the time series to obtain the biological clock sequence of the current sleep state, thereby obtaining the change trend of the sleep state, providing a data basis for the environmental optimization module, helping the system to adjust the sleep environment, and then optimize the sleep quality. The environment optimization module is electrically connected to the sleep device, and can adjust the parameters of the sleep device according to its sleep state, change trend and environmental data during sleep. The sleep aid unit determines whether the current sleep state is in the relaxation index according to the target biological data or dynamically adjusts the light mode and play mode of the sleep device based on the environmental data to assist sleep according to the change trend of the current sleep state. The awakening unit adjusts the light mode and play mode of the sleep device based on the change trend of the current sleep state, and changes the vibration mode of the sleep device according to the time of the change trend to assist in awakening the user. The sleep aid unit and the awakening unit improve the comfort and quality of sleep by precisely controlling the parameters of the sleep device. The data processing module is responsible for analyzing the historical adjustment data to determine whether the light mode adjusted by the current sleep device is correct. If an error occurs, the system re-determines the light mode of the sleep device through a clustering algorithm. In addition, the data processing module will establish a sleep optimization report based on the adjustment results of the sleep device, the target biological data and the change trend of the current sleep state. The sleep optimization report can help users understand their own sleep conditions, which is conducive to improving their own work and rest rules, thereby further improving sleep quality.

[0072] It is understandable that the system, by integrating sleep monitoring, sleep analysis, environmental optimization and data processing functions, can track the overall sleep process in real time and dynamically adjust the sleep environment to improve sleep quality. It adopts LSTM model, threshold logic algorithm and cluster analysis combined with intelligent environmental adjustment to achieve precise intervention in the sleep process, thereby improving the adaptability and scientificity of the system.

[0073] See also Figure 3 As shown, in some embodiments of the present application, when biological data and environmental data are collected according to a sensing device and the biological data is pre-processed to obtain target biological data, it includes:

[0074] Sensing devices include: wearable devices and sensors;

[0075] Collect biological data through wearable devices and environmental data through sensors;

[0076] Biological data include heart rate data, respiratory rate data and body movement times;

[0077] Environmental data include environmental humidity data and environmental light intensity;

[0078] The heart rate data and respiratory rate data are preprocessed, and the preprocessing includes data cleaning and standardization. The target heart rate and target respiratory rate are obtained based on the preprocessing results.

[0079] In some embodiments of the present application, when a threshold logic algorithm is used to derive a sleep state based on target biological data, the method includes:

[0080] The target heart rate is recorded as HR, the target respiratory rate is recorded as RR, and the number of body movements is recorded as BM;

[0081] When BM>5 or HR>80, the current sleep state is judged as awake;

[0082] When 15≤RR≤20, and 60≤HR≤75, the current sleep state is determined to be the first sleep state;

[0083] When RR < 15 and HR < 60, the current sleep state is determined to be the second sleep state;

[0084] When the awake state, the first sleep state or the second sleep state is not satisfied, the current sleep state is determined as an eye movement state.

[0085] Specifically, the sensing device includes a wearable device and a sensor. The user can continuously monitor his or her heart rate data, respiratory rate data and body movement number through the wearable device. The wearable device refers to a smart bracelet or smart watch or other forms of devices. The sensor includes a humidity sensor and a light sensor, and the sensor is responsible for obtaining the environmental data of the environment. The heart rate data and respiratory rate data are preprocessed. The purpose of data cleaning is to remove abnormal values ​​and noise that appear during the collection process. The target heart rate and target respiratory rate after standardization can be substituted into the threshold logic algorithm to avoid that different data cannot be compared and analyzed at the same scale. The target heart rate, target respiratory rate and body movement number reflect the physiological status during sleep. If BM is greater than 5 times or HR is greater than 80 times / minute, the system determines that the sleep state is awake. When RR is between 15 and 20 times / minute and HR is between 60 and 75 times / minute, the system will determine it as the first sleep state, which means a light sleep state.

[0086] When RR is less than 15 times / minute and HR is less than 60 times / minute, the system will judge it as the second sleep state, which indicates a deep sleep state. If the target heart rate, target breathing rate and body movement times do not meet the conditions of the above three states, the system will judge it to be in an eye movement state, indicating rapid eye movement sleep (REM).

[0087] It is understandable that by obtaining the target heart rate, target breathing rate and number of body movements in real time, the accuracy of sleep state analysis is improved, and the threshold logic algorithm can dynamically distinguish sleep states, thereby reflecting sleep quality and health status, laying a data foundation for subsequent adjustments to the environment to improve sleep quality.

[0088] In some embodiments of the present application, when the LSTM model is used to predict the biological clock sequence of the current sleep state according to the historical sleep state data, it includes:

[0089] Divide the historical sleep state data into a sleep training set and a sleep test set;

[0090] Use cross-validation combined with grid search to find the model parameters of the LSTM model, establish the LSTM model, fit the LSTM model based on the sleep training set, and substitute the sleep test set into the LSTM model to calculate the pass rate of predicting the biological clock sequence;

[0091] When the pass rate reaches the preset pass rate threshold, the current sleep state is substituted into the LSTM model to predict the biological clock sequence of the current sleep state.

[0092] In some embodiments of the present application, when the change trend of the current sleep state is obtained according to the biological clock sequence, it includes:

[0093] Extract the features of the biological clock sequence, including: the time of transition between different sleep states, the duration of sleep states, and the changing trend of sleep states;

[0094] The change trend includes a primary trend and a secondary trend. The primary trend is that the first sleep state is converted into the second sleep state, and the secondary trend is that the second sleep state is converted into the first sleep state.

[0095] Specifically, the historical sleep state data records the sleep state of different time periods. The historical sleep state data is divided into a sleep training set and a sleep test set. 60%-70% of the data is used as the sleep training set, and the rest is used as the sleep test set, ensuring that the sleep training set and the sleep test set can contain a variety of sleep state data in different time periods, thereby improving the generalization ability of the model. Cross-validation is used in combination with grid search to find the model parameters of the LSTM model. Cross-validation trains the LSTM model multiple times to verify its stability and performance. Grid search improves the accuracy and stability of the LSTM model by searching for the parameter combination of the LSTM model in the parameters and fitting the LSTM model with the sleep training set data. The sleep test set data is substituted into the trained LSTM model, and the pass rate of the model predicting the biological clock sequence is calculated. The pass rate reflects the performance of the LSTM model on unknown data and is an important benchmark for evaluating the performance of the LSTM model. After the model reaches the preset pass rate threshold, the current sleep state is substituted into the LSTM model to predict the biological clock sequence of the current sleep state, which improves the scientific nature of the prediction, reduces the errors of human judgment and prediction, and provides an efficient and accurate biological clock sequence.

[0096] It is understandable that the biological clock sequence reveals the overall sleep process and the sleep patterns therein. The time for transitioning between different sleep states indicates the time required to transition from one sleep state to another, which improves the accuracy of judging entering different sleep states, thus laying the foundation for subsequent adjustment of sleep equipment. The duration of the sleep state indicates the duration of each sleep state. By analyzing the duration of different sleep states, the time distribution of different sleep states can be obtained, thereby revealing the sleep state patterns therein. The changing trends of sleep states are divided into primary trends and secondary trends. The primary trend indicates the transition from light sleep to deep sleep, and the secondary trend indicates the transition from deep sleep to light sleep. By extracting and analyzing these features, the cycle of sleep states and changes in the biological clock can be accurately judged, laying data support for subsequent adjustment of sleep equipment, thereby improving sleep quality.

[0097] See also Figure 4 As shown, in some embodiments of the present application, when judging whether the current sleep state is in the relaxation index based on the target biological data or adjusting the light mode and the play mode of the sleep device based on the environmental data according to the change trend of the current sleep state, it includes:

[0098] Sleep equipment includes: sleep lights, sleep speakers and humidifiers;

[0099] When HR≤75 and BM≤3, the sleep state is judged as being in the relaxation index, otherwise the sleep state is not in the relaxation index;

[0100] When the sleep state is judged to be at the relaxation index or the change trend is the first level trend, the play mode of the sleep audio is adjusted to the white noise mode, the on / off state of the humidifier is adjusted according to the ambient humidity data, and the light mode of the sleep device is adjusted according to the ambient light intensity;

[0101] When the ambient humidity data is less than the set humidity threshold, the humidifier is turned on; when the ambient humidity data is greater than or equal to the set humidity threshold, the humidifier is not turned on;

[0102] When the ambient light intensity is greater than the set light threshold, the light mode of the sleep light is adjusted to the soft mode. When the ambient light intensity is less than the set light threshold, the light mode of the sleep light is adjusted to the fill light mode. When the ambient light intensity is equal to the set light threshold, the sleep light is not turned on.

[0103] Specifically, when the target heart rate and body movement times meet HR≤75 and BM≤3, the current sleep state is judged to be in the relaxation index. If this condition is not met, it means that the sleep state has not yet entered the relaxation index. In addition to adjusting the lighting mode and playback mode of the sleep device according to the target biological data, the lighting mode and playback mode of the sleep device are also adjusted based on the changing trend of the sleep state. When the changing trend is a first-level trend, it means a transition from a light sleep state to a deep sleep state, which is also a condition for adjusting the sleep device. When it is in the relaxation index or the changing trend is a first-level trend, the conditions for adjusting the sleep device are met. At this time, the playback mode of the sleep audio will be adjusted to the white noise mode. White noise is a uniform noise that can cover up external interference sounds, create a quiet sleep environment, and help accelerate falling asleep. The humidifier is turned on and off according to the ambient humidity data, and the light mode of the sleep device is adjusted according to the ambient light intensity. The humidity threshold can be adjusted according to actual needs. When the ambient humidity data is lower than the set humidity threshold, the humidifier will be turned on. Too low humidity will cause dry skin, respiratory discomfort and other problems, affecting the quality of falling asleep. When the ambient humidity reaches or exceeds the set humidity threshold, the system will turn off the humidifier to maintain a reasonable humidity range. The light threshold can be adjusted according to actual needs. When the ambient light intensity is higher than the set light threshold, the sleep light will be adjusted to a soft mode to avoid strong light stimulation and affect sleep quality. If the ambient light intensity is exactly equal to the set light threshold, the system will not turn on the sleep light to avoid interference from the light source. When the ambient light intensity is less than the set light threshold, the light mode of the sleep light is adjusted to the fill light mode. The fill light mode simulates the natural light changes from sunset to night, reducing the proportion of blue light and enhancing red light, effectively inhibiting the secretion of melatonin, thereby promoting the induction effect of sleep.

[0104] It is understandable that through the synergistic effect of white noise, humidifier and lighting modes, it is possible to effectively improve sleep quality, reduce the interference of environmental factors on sleep, create a suitable sleeping environment, thereby improving sleep continuity, and dynamically adjust sleep audio, humidifier and lighting modes, thereby improving the flexibility and automation of the system.

[0105] In some embodiments of the present application, when adjusting the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and changing the vibration mode of the sleep device according to the time of the change trend, it includes:

[0106] Sleeping equipment also includes: smart mattresses;

[0107] When the change trend is a secondary trend, the light mode of the sleep light is adjusted to a bright mode, the play mode of the sleep audio is adjusted to a light music mode, and the vibration mode of the smart mattress is adjusted to a first vibration mode;

[0108] According to the time of conversion of different sleep states, the exchange time of converting the second sleep state to the first sleep state is obtained. When the data processing module determines that the second sleep state is not converted to the first sleep state within the exchange time, the play mode of the sleep audio is adjusted to the heavy metal mode, and the vibration mode of the smart mattress is adjusted to the second vibration mode.

[0109] It can be understood that when the changing trend is a secondary trend, indicating a change from a deep sleep state to a light sleep state, the lighting mode of the sleep light is adjusted to a bright mode, and the playing mode of the sleep audio is adjusted to a light music mode, and the vibration mode of the smart mattress is adjusted to a first vibration mode to stimulate awakening. According to the exchange time of the second sleep state to the first sleep state, if it is judged that the second sleep state is not converted to the first sleep state within the exchange time, it indicates that the person is currently in a state of "staying in bed" and has not effectively gotten up. At this time, the system dynamically increases the stimulation intensity, adjusts the playing mode of the sleep audio to a heavy metal mode, and adjusts the vibration mode of the smart mattress to a second vibration mode. The vibration intensity of the second vibration mode is higher than that of the first vibration mode. The dynamic awakening mechanism improves the adaptability and automation of the system.

[0110] In some embodiments of the present application, when judging whether the light mode of the sleep device is correct according to the historical adjustment data, it includes:

[0111] The historical adjustment data includes historical sleep state data of the same type as the current sleep state, historical change trends, and corresponding historical adjustment light mode data;

[0112] The standardized result of adjusting the light mode according to the change trend of the current sleep state and the sleep state relaxation index is compared with the historical light mode adjustment data to obtain a ratio Q;

[0113] When Q is equal to 1, it means that the lighting mode is correct, and it is determined that the lighting mode of the sleeping light is not corrected;

[0114] When Q is not equal to 1, it indicates that the lighting mode is incorrect, and it is determined that the lighting mode of the sleep light should be corrected.

[0115] In some embodiments of the present application, if it is incorrect, a clustering algorithm is used to redetermine the light mode of the sleeping device.

[0116] Obtain representative data of each historically adjusted light mode data according to the historical adjustment data, and establish an aggregated data set by combining the representative data and the light mode of the current sleeping device;

[0117] Extract feature vectors of each data in the aggregated data set and determine important features;

[0118] Determine the expected number of clusters k as 3 and initialize the parameters of the Gaussian distribution;

[0119] Calculate the probability that each data in the aggregated data set belongs to each Gaussian distribution and obtain the responsibility value;

[0120] The cluster with the largest responsibility value is selected as the re-determined lighting mode of the current sleeping device.

[0121] Specifically, the lighting mode of the sleep light will be adjusted frequently, which may result in incorrect adjustment. Therefore, it is necessary to judge the lighting mode of the sleep light to avoid the lighting mode being turned on incorrectly and affecting the sleep quality. The same type of historical sleep state data, historical change trend and corresponding historical adjustment light mode data as the current adjustment light mode are obtained. For example, if the current change trend is a secondary trend, the light mode of the sleep light is adjusted to a bright mode. The historical change trend is a secondary trend, and the light mode of the sleep light is adjusted to a bright mode. The same type of historical adjustment light mode data should be consistent with the standardized result of the current adjustment light mode. The standardized result indicates that the dimension is consistent with the historical adjustment light mode data. When the ratio Q is equal to 1, it indicates that it is consistent with the historical adjustment light mode, the light mode is correct, and there is no need to correct the light mode of the sleep light. When Q is not equal to 1, it indicates that it is inconsistent with the historical adjustment light mode, indicating that the light mode is incorrect, and the light mode of the sleep light needs to be corrected.

[0122] It is understandable that using a clustering algorithm to redefine the light mode of the sleep device avoids the error caused by relying solely on preliminary judgments and improves the accuracy of the system's light mode classification. The Gaussian mixture model allows the system to automatically adjust the classification parameters based on the natural distribution of the data, thereby redefining the light mode of the current sleep device, improving the accuracy of the light mode adjustment, and improving the adjustment accuracy of the system.

[0123] To sum up, the beneficial effects of the present invention are: by collecting and analyzing biological data and environmental data, dynamically adjusting the sleep equipment, creating a suitable sleeping environment, and improving the quality of sleep, the sleep equipment is dynamically adjusted according to each user's biological data, sleep status and historical sleep data, and the biological clock sequence is predicted by the LSTM model, which can accurately grasp the biological rhythm, so as to appropriately adjust the sleep equipment, continuously track the trend of changes in sleep status, and monitor and adjust the mode of the sleep equipment in real time, which not only improves the sleep quality, but also can correct the sleep equipment according to historical adjustment data, further improves the accuracy of sleep optimization, effectively reduces interference factors, thereby improving the overall health level, and generates detailed sleep optimization reports, improves the comprehensive utilization of data, and enhances the adaptability and flexibility of the system.

[0124] In another preferred embodiment based on the above embodiment, refer to Figure 5 As shown, this embodiment provides an intelligent sleep state analysis based on multi-dimensional biological data, which is used to apply the above-mentioned dynamic optimization system based on multi-dimensional biological data, including:

[0125] S100: collecting biological data and environmental data, and preprocessing the biological data to obtain target biological data;

[0126] S200: deriving the sleep state by using a threshold logic algorithm based on the target biological data, predicting the biological clock sequence of the current sleep state by using an LSTM model based on the historical sleep state data, and deriving the change trend of the current sleep state based on the biological clock sequence;

[0127] S300: dynamically adjusting the sleep device, judging whether the current sleep state is in the relaxation index based on the target biological data or adjusting the light mode and the play mode of the sleep device based on the environmental data according to the change trend of the current sleep state, adjusting the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and changing the vibration mode of the sleep device according to the time of the change trend;

[0128] S400: Determine whether the light mode of the sleep device is correct based on historical adjustment data. If it is not correct, use a clustering algorithm to re-determine the light mode of the sleep device, and create a sleep optimization report based on the adjustment results of the sleep device, target biological data, and the change trend of the current sleep state.

[0129] It is understandable that in S100, biological data and environmental data are collected, and biological data is preprocessed, which improves the data quality of the target biological data and lays the foundation for the subsequent analysis of the target biological data. In S200, the threshold logic algorithm and LSTM model predict the sleep state and biological clock sequence. Based on the target biological data, the threshold logic algorithm is used to make a preliminary judgment on the sleep state, and the LSTM model is used to predict the current biological clock sequence using the historical sleep state data. By analyzing the biological clock sequence, the change trend of the current sleep state is obtained, which provides data support for the subsequent adjustment of the sleep device. In S300, according to the target biological data and the change trend of the current sleep state, the sleep device is dynamically adjusted, and it is judged whether the current sleep state is in the relaxation index, or the lighting mode and the playback mode are dynamically adjusted based on the environmental data according to the change trend. It is conducive to improving the surrounding environment and thus improving sleep quality. S400 redefines the lighting mode of the sleep device and generates an optimization report. It verifies whether the lighting mode of the sleep device is correctly turned on based on historical adjustment data, readjusts the lighting mode of the sleep device through a clustering algorithm, and combines the current biological data and changing trends of the sleep state to generate a personalized sleep optimization report. It can intelligently judge and adjust the sleep device, improve sleep quality, and thus improve sleep health.

[0130] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0131] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0132] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0133] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A dynamic optimization system based on multidimensional biological data, characterized in that: include: Sleep monitoring module, sleep analysis module, environment optimization module and data processing module; The sleep monitoring module is configured to be electrically connected to the sensing device, collect biological data and environmental data according to the sensing device, and pre-process the biological data to obtain target biological data; The sleep analysis module is configured to use a threshold logic algorithm to obtain a sleep state based on the target biological data, use an LSTM model to predict a biological clock sequence of a current sleep state based on historical sleep state data, and obtain a change trend of the current sleep state based on the biological clock sequence; The environment optimization module is configured to dynamically adjust the sleep device, the environment optimization module is electrically connected to the sleep device, and the environment optimization module includes a sleep aid unit and a wake-up unit, the sleep aid unit determines whether the current sleep state is within the relaxation index based on the target biological data or adjusts the light mode and the play mode of the sleep device based on the environment data according to the change trend of the current sleep state, the wake-up unit adjusts the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and changes the vibration mode of the sleep device according to the time of the change trend; The data processing module is configured to determine whether the lighting mode of the sleep device is correct based on historical adjustment data. If it is incorrect, a clustering algorithm is used to re-determine the lighting mode of the sleep device, and a sleep optimization report is established based on the adjustment results of the sleep device, target biological data, and the changing trend of the current sleep state.

2. The multi-dimensional biological data dynamic optimization system according to claim 1, characterized in that: When the biological data and environmental data are collected according to the sensing device and the biological data are pre-processed to obtain the target biological data, it includes: The sensing device includes: a wearable device and a sensor; Collect the biological data through the wearable device, and collect the environmental data through the sensor; The biological data includes heart rate data, respiratory rate data and body movement times; The environmental data includes environmental humidity data and environmental light intensity; The heart rate data and the respiratory rate data are preprocessed, wherein the preprocessing includes data cleaning and standardization, and a target heart rate and a target respiratory rate are obtained based on the results of the preprocessing.

3. The multi-dimensional biological data dynamic optimization system according to claim 2, characterized in that: When the sleep state is derived based on the target biological data using a threshold logic algorithm, it includes: The target heart rate is recorded as HR, the target respiratory rate is recorded as RR, and the number of body movements is recorded as BM; When BM>5 or HR>80, the current sleep state is judged as awake; When 15≤RR≤20, and 60≤HR≤75, the current sleep state is determined to be the first sleep state; When RR < 15 and HR < 60, the current sleep state is determined to be the second sleep state; When the awake state, the first sleep state or the second sleep state is not satisfied, the current sleep state is determined as an eye movement state.

4. The multi-dimensional biological data dynamic optimization system according to claim 3, characterized in that: When using the LSTM model to predict the biological clock sequence of the current sleep state based on historical sleep state data, it includes: Dividing the historical sleep state data into a sleep training set and a sleep test set; Using cross validation combined with grid search to find model parameters of the LSTM model, establishing the LSTM model, fitting the LSTM model according to the sleep training set, and substituting the sleep test set into the LSTM model to calculate the pass rate of predicting the biological clock sequence; When the pass rate reaches a preset pass rate threshold, the current sleep state is substituted into the LSTM model to predict the biological clock sequence of the current sleep state.

5. The multi-dimensional biological data dynamic optimization system according to claim 4, characterized in that: When the change trend of the current sleep state is derived according to the biological clock sequence, it includes: Extracting features of the biological clock sequence, the features including: the time of transition between different sleep states, the duration of the sleep state, and the change trend of the sleep state; The change trend includes a primary trend and a secondary trend, the primary trend is that the first sleep state is converted into the second sleep state, and the secondary trend is that the second sleep state is converted into the first sleep state.

6. The multi-dimensional biological data dynamic optimization system according to claim 5, characterized in that: When judging whether the current sleep state is within the relaxation index based on the target biological data or adjusting the light mode and the play mode of the sleep device based on the environmental data according to the change trend of the current sleep state, it includes: The sleeping equipment includes: sleeping lights, sleeping speakers and humidifiers; When HR≤75 and BM≤3, the sleep state is judged as being in the relaxation index, otherwise the sleep state is not in the relaxation index; When it is determined that the sleep state is at the relaxation index or the change trend is a primary trend, the play mode of the sleep sound is adjusted to a white noise mode, the on / off state of the humidifier is adjusted according to the ambient humidity data, and the light mode of the sleep device is adjusted according to the ambient light intensity; When the ambient humidity data is less than the set humidity threshold, the humidifier is turned on; when the ambient humidity data is greater than or equal to the set humidity threshold, the humidifier is not turned on; When the ambient light intensity is greater than the set light threshold, the light mode of the sleep light is adjusted to the soft mode. When the ambient light intensity is less than the set light threshold, the light mode of the sleep light is adjusted to the fill light mode. When the ambient light intensity is equal to the set light threshold, the sleep light is not turned on.

7. The multi-dimensional biological data dynamic optimization system according to claim 6, characterized in that: When adjusting the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and changing the vibration mode of the sleep device according to the time of the change trend, it includes: The sleeping device also includes: a smart mattress; When the change trend is a secondary trend, the light mode of the sleep light is adjusted to a bright mode, the play mode of the sleep audio is adjusted to a light music mode, and the vibration mode of the smart mattress is adjusted to a first vibration mode; According to the time of conversion of the different sleep states, the exchange time of converting the second sleep state to the first sleep state is obtained, and when the data processing module determines that the second sleep state is not converted to the first sleep state within the exchange time, the play mode of the sleep sound is adjusted to the heavy metal mode, and the vibration mode of the smart mattress is adjusted to the second vibration mode.

8. The multi-dimensional biological data dynamic optimization system according to claim 7, characterized in that: When judging whether the light mode of the sleep device is correct according to the historical adjustment data, it includes: The historical adjustment data includes historical sleep state data of the same type as the current sleep state, historical change trends, and corresponding historical adjustment light mode data; The standardized result of adjusting the light mode according to the change trend of the current sleep state and the sleep state relaxation index is compared with the historical light mode adjustment data to obtain a ratio Q; When Q is equal to 1, it indicates that the light mode is correct, and it is determined that the light mode of the sleep light is not to be modified; When Q is not equal to 1, it indicates that the lighting mode is incorrect, and it is determined that the lighting mode of the sleep light should be corrected.

9. The multi-dimensional biological data dynamic optimization system according to claim 8, characterized in that: If it is incorrect, a clustering algorithm is used to redetermine the light mode of the sleeping device. Obtain representative data of each historically adjusted light mode data according to the historical adjustment data, and establish an aggregated data set by combining the representative data and the light mode of the current sleeping device; Extracting feature vectors of each data in the aggregated data set and determining important features; Determine the expected number of clusters k as 3 and initialize the parameters of the Gaussian distribution; Calculating the probability that each data in the aggregated data set belongs to each Gaussian distribution to obtain a responsibility value; The cluster with the largest responsibility value is selected as the re-determined lighting mode of the current sleeping device.

10. An intelligent sleep state analysis based on multidimensional biological data, used for applying the multidimensional biological data based dynamic optimization system according to any one of claims 1 to 9, characterized in that: include: Collecting biological data and environmental data, and preprocessing the biological data to obtain target biological data; Based on the target biological data, a threshold logic algorithm is used to obtain the sleep state, and a LSTM model is used to predict the biological clock sequence of the current sleep state according to the historical sleep state data, and a change trend of the current sleep state is obtained according to the biological clock sequence; Dynamically adjust the sleep device, determine whether the current sleep state is in the relaxation index based on the target biological data, or adjust the light mode and the play mode of the sleep device based on the environmental data according to the change trend of the current sleep state, adjust the light mode and the play mode of the sleep device based on the change trend of the current sleep state, and change the vibration mode of the sleep device according to the time of the change trend; Whether the lighting mode of the sleep device is correct is determined based on historical adjustment data. If incorrect, a clustering algorithm is used to re-determine the lighting mode of the sleep device. A sleep optimization report is established based on the adjustment results of the sleep device, target biological data, and the changing trend of the current sleep state.

Images