Intelligent infant monitoring system and method

Abstract

The invention discloses a baby intelligent monitoring system and method, and relates to the technical field of baby intelligent monitoring, a trained environment automatic control model is used to adjust key factors affecting baby sleep quality, and a sleep health report is constructed according to baby sleep data; carrying out anomaly detection on the collected baby state data, starting a multi-stage alarm mechanism and sending out a multi-stage alarm instruction when an abnormal event exists, and selecting and switching each guardian when sending out the alarm instruction according to the position data and response data of the guardian; the alarm instruction and the matched emergency plan are sent to the selected guardian and confirmed by the guardian, and the alarm sensitivity and the guardian receiving the alarm instruction are adjusted according to the infant state condition data and the alarm feedback data. And a confirmation mechanism and a channel switching mechanism are introduced, so that the risk that a guardian misses all alarm instructions is reduced, and the infant can be timely handled when the infant is abnormal.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent infant monitoring, and in particular to an intelligent infant monitoring system and method. Background Art

[0002] The smart baby monitoring system integrates high-definition cameras, precision sensors, artificial intelligence analysis algorithms and convenient mobile Internet technology, providing new parents with a comprehensive, all-weather baby monitoring solution. It can capture the baby's every move in real time, and through high-definition video and audio transmission, parents can check the baby's status at any time no matter where they are. It can also use built-in sensors to accurately monitor important physiological and environmental indicators of the baby's breathing rate, heart rate changes, body temperature fluctuations, and sleeping environment (such as indoor temperature, humidity and air quality).

[0003] With advanced intelligent analysis technology, the baby smart monitoring system can automatically identify the baby's emotional changes, sleep patterns and even potential abnormal conditions, such as sudden crying, high or low body temperature, and send an alarm notification to the parents' mobile phones immediately, ensuring that parents can respond quickly and give the baby timely love and care. In addition, many smart monitoring systems are also equipped with growth recording and analysis functions to help parents track the baby's growth trajectory and understand its behavior patterns, so as to participate in the parenting process more scientifically and efficiently.

[0004] In the Chinese invention patent with application publication number CN107146375A, a method and system for intelligently monitoring the sleeping state of infants and young children is disclosed, which is applied to the monitoring terminal device set at the end of the monitored object, and at least includes the following steps: S10 continuously obtains the image of the monitored object; S20 determines whether the monitored object is located in a preset dangerous area according to the image; S30 sends a reminder signal to the mobile terminal when the monitored object is determined to be located in the dangerous area; S40 detects the distance between the mobile terminal carried by the guardian and the bed; S50 controls the safety protection component set at the side of the bed to start protection when it is detected that the distance between the mobile terminal and the bed is greater than the preset unsafe distance threshold. Determine the distance between the guardian and the monitored object. When the distance is far, control the safety protection component set at the side of the bed to start protection to avoid danger, thereby further ensuring the safety of the monitored object.

[0005] Combined with the above application and the contents of the prior art: When the baby is sleeping or moving alone, the baby needs to be monitored and supervised in real time, such as paying attention to the baby's sleep quality, body temperature and movement status. When the baby has abnormalities in the monitoring area, an alarm is sent to the outside so that the guardian can deal with it in time.

[0006] When the guardian is not around or even at a certain distance from the monitoring area, existing remote monitoring methods usually monitor the baby's various data in real time when monitoring the baby, and identify abnormal data, abnormal behavior and abnormal events related to the baby in real time. For example, when the baby's body temperature is abnormal or bumped, the guardian is guided to deal with it. However, in such scenarios, especially when the environmental conditions in the monitoring area are poor and not suitable for the baby's sleep and activities, the probability of the baby having abnormalities is also relatively high. If a single alarm command is still issued when the baby has abnormalities, there is a certain risk of the guardian ignoring or even missing it, which may have serious consequences.

[0007] To this end, the present invention provides a baby intelligent monitoring system and method. Summary of the invention

[0008] 1. Technical issues to be resolved In view of the deficiencies of the prior art, the present invention provides an intelligent infant monitoring system and method, which uses a trained environmental automatic control model to adjust key factors affecting infant sleep quality, and constructs a sleep health report based on infant sleep data; performs abnormal detection on the collected infant status data, enables a multi-level alarm mechanism and issues a multi-level alarm instruction when an abnormal event occurs, and selects and switches each guardian when issuing an alarm instruction based on the guardian's location data and response data; issues an alarm instruction and a corresponding emergency plan to the selected guardian and is confirmed by the guardian, and adjusts the alarm sensitivity and the guardian who receives the alarm instruction based on the infant status condition data and alarm feedback data. The introduction of a confirmation mechanism and a channel switching mechanism reduces the risk of guardians missing all alarm instructions, thereby solving the technical problems recorded in the background technology.

[0009] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for intelligent monitoring of infants, comprising: collecting environmental data of a monitoring area by a sensor network, and constructing an environmental coefficient from continuous environmental data , if the environmental factor If the environmental threshold is lower than the threshold, a monitoring instruction will be issued to the outside world; After collecting the baby's sleep data, the sleep quality is evaluated. When the sleep quality is lower than expected, the trained automatic environmental control model is used to adjust the key factors affecting the baby's sleep quality, and a sleep health report is constructed based on the baby's sleep data; Perform abnormal detection on the collected infant status data, activate the multi-level alarm mechanism and issue multi-level alarm instructions when abnormal events occur, and switch the current alarm channel if the alarm instruction is not responded to in time; If there are multiple abnormal events, the alarm priority of each abnormal event is determined according to the degree of abnormality, and each guardian is selected and switched when issuing an alarm command based on the guardian's location data and response data; The alarm instruction and the corresponding emergency plan are sent to the selected guardian and confirmed by the guardian. According to the infant's status condition data and alarm feedback data, the alarm sensitivity and the guardian who receives the alarm instruction are adjusted; among them, the warning threshold The adjustment is as follows: ; Where: is the basic value of the warning threshold, In the time window Any time node, k is the sleep adjustment coefficient, For the time window Internal sleep quality value Below quality threshold The number of times, is the rate of change of environmental conditions, is the rate of change adjustment factor.

[0010] Furthermore, the sensor network arranged in the monitoring area collects environmental data in the monitoring area and aggregates them to generate an environmental data set; the environmental data in the monitoring area is used as input, and the environmental conditions in the monitoring area are evaluated using the trained environmental evaluation model to obtain a corresponding environmental score.

[0011] Furthermore, the environmental scores are arranged along the time axis to construct the environmental coefficient. , as follows: Score the acquired environment Perform linear normalization and project the corresponding data values ​​onto According to the following formula: ; in, , The number of environmental scores, , ,and , For the i Environmental ratings, is the mean value of the environmental score, The qualified standard value for environmental scoring.

[0012] Furthermore, a monitoring device located in the monitoring area records the sleep quality data of the infant in the monitoring area; the sleep quality data is used as input, and the trained quality evaluation model is used to evaluate the sleep quality of the infant to obtain a sleep quality value; Principal component analysis is used to screen out the key factors that affect the infant's sleep quality from various environmental parameters in the monitoring area, and the corresponding degree of influence is determined; when the infant's sleep quality is lower than expected, the trained environmental automatic control model is used to prioritize the adjustment and optimization of the key factors.

[0013] Furthermore, after a preset observation period, a sleep health report is constructed based on the infant's sleep data, and the reporting interval of the sleep health report is constrained. , the constraints are as follows: ; in, b is the number of sleep quality values, ; It is i The sleep quality value and j The difference between the sleep quality values, is the average of the differences, , ,and .

[0014] Furthermore, the monitoring equipment located in the monitoring area collects the monitoring parameters of the infant, including sleep quality data, activity status data and vital sign data; the infant monitoring parameters are used as input and the trained anomaly detection model is used for detection. When an abnormal event is detected, a first-level alarm instruction is issued to the outside.

[0015] Furthermore, after multiple first-level alarm instructions are issued continuously, under dimensionless conditions, the warning value is constructed based on the issuance status data of the first-level alarm instructions. , if the warning value obtained When the preset alarm threshold is exceeded, a secondary alarm command is issued to the outside; the alarm value is constructed based on the status data of the primary alarm command. The way is as follows: ; Where: For the i Second and i +1 time interval between alarms, For the i The abnormality level at the time of the alarm, and are weight coefficients, , , is the attenuation coefficient of the cumulative impact, .

[0016] Further, after detecting that the infant's state is abnormal, the abnormal data of the abnormal event is used as input, and the trained abnormal evaluation model is used to perform abnormal evaluation and obtain the abnormal value; Each abnormal event is marked with an abnormal value, and after sorting each abnormal event according to the size of the abnormal value, the corresponding alarm priority is obtained.

[0017] Further, the location information of each guardian is obtained to obtain the distance data between the guardian and the baby, as well as the response speed of each guardian after receiving the alarm command; Responsiveness constructed from distance data and response speed , according to the responsiveness Sort each guardian from small to large and obtain the corresponding alarm ranking.

[0018] Furthermore, the distance and response speed Perform linear normalization and project the corresponding data values ​​onto The response is generated according to the following formula : ; Weight coefficient: , ,and , , The number of response speeds ， For the i The response speed, is the mean response speed, Its expected value.

[0019] Furthermore, emergency response plans are formulated for each abnormal event. After an abnormal event occurs and an alarm instruction is issued to the outside, the alarm instruction and the corresponding emergency plan are sent to the selected guardian and confirmed by the guardian; If the confirmation time exceeds the preset time, resend or switch the guardian; collect the corresponding alarm feedback data. If the guardian fails to confirm the alarm command twice in a row after receiving it, the guardian will no longer receive the alarm command.

[0020] Furthermore, after a preset adjustment period, the warning threshold is adjusted according to the environmental status and sleep quality. Adjustments are made to improve the alarm sensitivity as the optimization goal, and the qualified standard values ​​of various monitoring parameters are optimized according to the alarm feedback data using a pre-trained genetic algorithm to obtain the optimized qualified standard values.

[0021] An intelligent infant monitoring system includes an environmental analysis unit, which collects environmental data of the monitoring area through a sensor network and constructs an environmental coefficient based on the continuous environmental data. , if the environmental factor If the environmental threshold is lower than the threshold, a monitoring instruction will be issued to the outside world; The environmental condition adjustment unit collects the infant's sleep data and evaluates the sleep quality. When the sleep quality is lower than expected, the trained environmental automatic control model is used to adjust the key factors affecting the infant's sleep quality, and a sleep health report is constructed based on the infant's sleep data. The abnormality detection unit detects abnormalities in the collected infant status data, activates a multi-level alarm mechanism and issues a multi-level alarm command when an abnormal event occurs, and switches the current alarm channel if the alarm command is not responded to in time; Abnormal alarm unit, if there are multiple abnormal events in the current connection, the alarm priority of each abnormal event is determined according to the degree of abnormality, and each guardian is selected and switched when issuing an alarm command according to the guardian's location data and response data; The feedback control unit sends the alarm instruction and the corresponding emergency plan to the selected guardian for confirmation. According to the infant status condition data and the alarm feedback data, the alarm sensitivity and the guardian who receives the alarm instruction are adjusted.

[0022] (III) Beneficial effects The present invention provides an intelligent infant monitoring system and method, which have the following beneficial effects:

[0023] 1. Based on environmental coefficient It can comprehensively evaluate the environmental conditions in the monitoring area, adjust the environmental conditions in time, and start the automatic monitoring mechanism to prevent the health or condition of the baby from being affected in abnormal environments.

[0024] 2. Automatically evaluate the baby's sleep quality while the baby is sleeping. When the baby's sleep quality is poor, the guardian can make timely adjustments.

[0025] 3. Analyze and obtain the key factors that have a greater impact on the baby's sleep quality, use the trained environmental automatic control model to control the key factors, and control the environmental conditions in the monitoring area. This will serve as feedback to adjust the baby's sleep quality.

[0026] 4. By restricting the frequency of generating sleep health reports, when the sleep quality is poor, the frequency of sleep health reports can be increased. By sending the sleep quality report to the guardian, the guardian can adjust the baby's sleep quality in the current stage in real time, and make timely adjustments if there are any abnormalities.

[0027] 5. By detecting abnormal conditions, it is determined whether there is an abnormal event at present. When an abnormal event occurs, an alarm command is issued to the outside. Different from issuing a sleep health report, the guardian can realize remote monitoring.

[0028] 6. When the first-level alarm instruction fails to receive a timely response, an alert value is established, and then a second-level alarm instruction is issued as a feedback and response supplement to the first-level alarm instruction to prevent the first-level alarm instruction from being ignored; a confirmation mechanism and a channel switching mechanism are introduced for the second-level alarm instruction to reduce the risk of guardians missing all alarm instructions, so that when monitoring the baby, the baby can be dealt with in a timely manner when there are abnormalities or abnormal events.

[0029] 7. Responsiveness By sorting the guardians, you can select and switch the guardians. When the current abnormal event is more serious, you can choose the guardian who handles the abnormal event to ensure that the current abnormal event can be handled in time and the baby is protected when the baby is monitored.

[0030] 8. When sending an alarm command to the outside, cooperate with sending an emergency plan to alleviate the current abnormal state, switch and select the current guardians, and further avoid being ignored when the baby has an abnormality; adjust the warning threshold based on the feedback data, and adjust the alarm sensitivity according to the baby's current state, which can reduce the number of false alarms and invalid alarms. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a flow chart of the intelligent infant monitoring method of the present invention; Figure 2 The figure is a schematic diagram of the structure of the intelligent infant monitoring system of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] See also Figure 1 The present invention provides an intelligent infant monitoring method, comprising: Step 1: Collect environmental data of the monitoring area through the sensor network, and construct environmental coefficients from continuous environmental data , if the environmental factor If the environmental threshold is lower than the threshold, a monitoring instruction will be issued to the outside world; The step 1 includes the following contents: Step 101: After determining the sleeping and activity areas of the infant, the areas are used as monitoring areas, and a sensor network, such as a temperature and humidity sensor, an air quality sensor, and a noise sensor, is arranged in the monitoring area; the sensor network collects environmental data in the monitoring area, including temperature and humidity data, air quality data, white noise data, and light data, and generates an environmental data set by aggregating the data; The machine learning algorithm is trained with the labeled sample data to obtain a trained environmental evaluation model; the environmental data in the monitoring area is used as input, and the trained environmental evaluation model is used to evaluate the environmental conditions in the monitoring area and obtain the corresponding environmental score; When in use, the sensor network detects various data in the monitoring area, and evaluates the degree of environmental abnormality at the current time node based on the acquired environmental data, and can preliminarily determine whether the monitoring area is suitable for infant activities or sleep; Step 102: Arrange the environmental scores obtained at several consecutive time nodes along the time axis to construct an environmental coefficient , the environmental conditions in the monitoring area are evaluated according to the environmental coefficient, as follows: the environmental score obtained Perform linear normalization and project the corresponding data values ​​onto According to the following formula: ; in, , The number of environmental scores, , ,and , the weight coefficient can be obtained by referring to the hierarchical method; For the i Environmental ratings, is the mean value of the environmental score, The pass standard value for scoring the environment; Pre-set environmental thresholds based on historical data and management expectations for condition data in the monitoring area; If the environmental factor If the condition is lower than the environmental threshold, it means that the environmental conditions in the current monitoring area are poor, which may have a certain impact on the baby's sleep quality or activity status. In order to compensate or reduce this impact, the baby's current status needs to be observed and adjusted in real time. At this time, a monitoring instruction is issued to the outside. When using, combine the contents in steps 101 and 102: By constructing the corresponding environmental coefficient based on several consecutive environmental scores , according to the environmental factor A comprehensive evaluation can be made on the environmental conditions in the monitoring area. If the environmental conditions in the monitoring area may have a certain negative impact on the baby, the environmental conditions can be adjusted in time and the automatic monitoring mechanism can be activated to prevent the baby's health or condition from being affected in an abnormal environment.

[0034] When the guardian is not around or even at a certain distance from the monitoring area, existing remote monitoring methods usually monitor the baby's various data in real time when monitoring the baby, and identify abnormal data, abnormal behavior and abnormal events related to the baby in real time. For example, when the baby's body temperature is abnormal or bumps, the guardian is guided to deal with it. However, in such scenarios, especially when the environmental conditions in the monitoring area are poor and not suitable for the baby's sleep and activities, the probability of the baby having abnormalities is also relatively high. If a single alarm command is still issued when the baby has an abnormality, there is a certain risk of the guardian ignoring or even missing it, which may have serious consequences.

[0035] Step 2: After collecting the baby's sleep data, the sleep quality is evaluated. When the sleep quality is lower than expected, the trained automatic environmental control model is used to adjust the key factors affecting the baby's sleep quality, and a sleep health report is constructed based on the baby's sleep data; The step 2 includes the following contents: Step 201: A monitoring device located in the monitoring area, such as a smart wearable device, an image acquisition device, and a health monitoring device, records the sleep quality data of the infant such as the sleeping time, the number of tossing and turning, and the activity frequency in the monitoring area; The machine learning algorithm is trained with the labeled sample data to obtain a trained sleep quality evaluation model; the sleep quality data is used as input, and the trained quality evaluation model is used to evaluate the infant's sleep quality to obtain a sleep quality value; When in use, the sleep quality value is generated by collecting the baby's sleep quality data, and the baby's sleep quality is automatically evaluated when the baby is sleeping. When the baby's sleep quality is poor, the guardian can make timely adjustments; Step 202: Use principal component analysis to select key factors that have a greater impact on the infant's sleep quality from various environmental parameters in the monitoring area, and determine the corresponding impact degree; After training the machine learning algorithm with the labeled sample data, the trained environmental automatic control model is obtained; when the infant's sleep quality is lower than expected, the trained environmental automatic control model is used to adjust and optimize the key factors in priority; When used, as a further content, when the baby's sleep is poor, in order to improve the baby's sleep quality, the key factors that have a greater impact on the baby's quality are analyzed based on a number of historical data, and the trained environmental automatic control model is used to control the key factors, which can realize the control of the environmental conditions in the monitoring area, and use this as feedback to adjust the baby's sleep quality; Step 203: after a preset observation period, construct a sleep health report based on the infant's sleep data and constrain the reporting interval of the sleep health report , the constraints are as follows: ; in, b is the number of sleep quality values, ; It is i The sleep quality value and j The difference between the sleep quality values, is the average of the differences, , ,and , the value of the weight coefficient remains consistent with the previous value; When using, combine the contents in steps 201 to 203: By restricting the frequency of generating sleep health reports, the frequency of sleep health reports can be increased when the sleep quality is poor. By sending the sleep quality reports to guardians, guardians can adjust the sleep quality of the baby in the current stage in real time. If there are any abnormalities, adjustments can be made in time.

[0036] Step 3: Perform abnormality detection on the collected infant status data, enable a multi-level alarm mechanism and issue a multi-level alarm command when an abnormal event occurs, and switch the current alarm channel if the alarm command is not responded to in time; The step three includes the following contents: Step 301: Train a machine learning algorithm using the labeled sample data to obtain a trained anomaly detection model; The monitoring equipment located in the monitoring area collects the monitoring parameters of the baby, including sleep quality data, activity status data and vital sign data; Taking the infant monitoring parameters as input, the trained anomaly detection model is used to detect anomalies in the infant's sleeping and non-sleeping states. When an abnormality is detected, a first-level alarm command is issued to the outside. When in use, the baby's status data is continuously monitored and collected when the baby is sleeping or ends the current sleeping state, and by detecting abnormal states, it is determined whether there is an abnormal event at present. When there is an abnormal event, an alarm instruction is issued to the outside, which is different from issuing a sleep health report. The guardian can realize remote monitoring; Step 302: after multiple first-level alarm instructions are issued continuously, a multi-level alarm mechanism is established when no response is obtained. At this time, under dimensionless conditions, an alert value is established based on the issuance status data of the first-level alarm instruction. The way is as follows: ; Where: For the i Second and i +1 time interval between alarms, For the i The abnormality level at the time of the alarm, and are weight coefficients, , , is the attenuation coefficient of the cumulative impact, , The value can be 2.713; Pre-set warning thresholds based on historical data and expectations for infant care management; If the warning value obtained If the alarm exceeds the preset threshold, it means that the baby is in a serious abnormal state and needs to be treated in time. At this time, a secondary alarm command is issued to the outside. When used, an alert value is established when the first-level alarm command fails to receive a timely response , and then issue a second-level alarm command as a feedback and response supplement to the first-level alarm command to avoid the first-level alarm command being ignored; Step 303: If the secondary alarm instruction sent to the outside is not confirmed, the current alarm channel is switched to ensure that a backup system can take over the monitoring and alarm functions, for example, by sending push notifications, text messages or voice reminders through a mobile phone application, so as to ensure that the alarm can be timely notified even if the baby is not in the baby's room; When using, combine the contents in steps 301 to 303: As a further supplement, considering that the first-level or second-level alarm instructions may be ignored, a confirmation mechanism and a channel switching mechanism are introduced for the second-level alarm instructions to reduce the risk of guardians missing all alarm instructions, so that when monitoring the baby, the baby can be dealt with in a timely manner when there are abnormalities or abnormal events.

[0037] Step 4: If there are multiple abnormal events in the current connection, determine the alarm priority of each abnormal event according to the degree of abnormality, and select and switch each guardian when issuing an alarm instruction based on the guardian's location data and response data; The step 4 includes the following contents: Step 401: training a machine learning algorithm with the labeled sample data to obtain a trained anomaly evaluation model; After detecting that the infant's state is abnormal, the abnormal data of the abnormal event is used as input, and the trained abnormal evaluation model is used to perform abnormal evaluation and obtain the abnormal value; Each abnormal event is marked with an abnormal value, and after sorting each abnormal event according to the size of the abnormal value, the corresponding alarm priority is obtained. For example, abnormal vital signs such as apnea and abnormal heart rate are set as the highest priority, and temperature changes or activity alarms are set as the second priority, etc.; When in use, as a further content, it is considered that there are many types of abnormal events and the importance of each abnormal event is inconsistent. Therefore, if there are multiple abnormal events at the same time and an alarm is issued to the outside, each abnormal event can be processed in turn according to the importance level to reduce the probability of false alarms.

[0038] Step 402: when there are several guardians who receive the alarm instruction, obtain the location information of each guardian, obtain the distance data between the guardian and the baby, and the response speed of each guardian after receiving the alarm instruction; Responsiveness constructed from distance data and response speed , where the distance and response speed Perform linear normalization and project the corresponding data values ​​onto According to the following formula: ; Weight coefficient: , ,and , its specific value can be obtained by referring to the hierarchical analysis method; , The number of response speeds ， For the i The response speed, is the mean response speed, is its qualified expected value; Responsiveness After marking each guardian, sort each guardian according to the responsiveness from small to large and obtain the corresponding alarm ranking, and select and switch the guardian who receives the alarm instruction according to the alarm ranking; for example, the guardian with the highest alarm ranking will issue the alarm instruction first; When using, combine the contents in steps 401 and 402: As a further content, considering that there are many guardians receiving alarm instructions, and most abnormal events are actually easy to handle, in order to avoid affecting all guardians, a responsiveness is constructed based on historical data. By sorting the guardians, you can select and switch the guardians. At the same time, considering the distance and response level of the guardians, when the current abnormal event is more serious, you can choose the guardian who handles the abnormal event to ensure that the current abnormal event can be handled in time and the baby is protected when the baby is monitored.

[0039] Step 5: Send the alarm instruction and the corresponding emergency plan to the selected guardian and have the guardian confirm it. Adjust the alarm sensitivity and the guardian who receives the alarm instruction according to the infant's status condition data and the alarm feedback data. The step five includes the following contents: Step 501: formulate clear emergency response plans for each abnormal event, such as abnormal plans for apnea, suffocation risk and abnormal body temperature; after an abnormal event occurs and an alarm instruction is issued to the outside, the alarm instruction and the corresponding emergency plan are issued to the selected guardian and confirmed by the guardian; If the confirmation time exceeds the preset time, resend or switch the guardian; collect corresponding alarm feedback data, such as alarm reliability and timeliness data, subsequent detection and diagnosis data, etc.; among them, if the guardian fails to confirm the alarm instruction twice in a row after receiving it, he will no longer receive the alarm instruction; As further feedback, when an abnormal event occurs with the baby, in order to ensure that the abnormal event can be handled in a targeted manner, an emergency plan is sent in conjunction with the external alarm command. When used as a reference, the current abnormal state can be alleviated. As a further processing method, several current guardians are switched and selected to further avoid being ignored when the baby has an abnormality.

[0040] Step 502: After a preset adjustment period, adjust the warning threshold according to the environmental status and sleep quality. Adjust the alarm sensitivity as follows: ; Where: is the basic value of the warning threshold, In the time window At any time node, k is the sleep adjustment coefficient, which ranges from 0 to 1; is an indicator function, which takes the value of 1 when the condition is true; For the time window Internal sleep quality value Below quality threshold The number of times, is the rate of change of environmental conditions, is the change rate adjustment coefficient, which ranges from 0 to 1; With the improvement of alarm sensitivity as the optimization goal, the qualified standard values ​​of various monitoring parameters are optimized according to the alarm feedback data using the pre-trained genetic algorithm to obtain the optimized qualified standard values; When using, combine the contents in steps 501 and 502: As a further improvement, by adjusting the warning threshold based on the feedback data and adjusting the alarm sensitivity according to the current state of the baby, the number of false alarms and invalid alarms can be reduced.

[0041] See also Figure 2 The present invention provides an intelligent infant monitoring system, comprising: Environmental analysis unit, which collects environmental data of the monitoring area through the sensor network and constructs environmental coefficients based on continuous environmental data , if the environmental factor If the environmental threshold is lower than the threshold, a monitoring instruction will be issued to the outside world; The environmental condition adjustment unit collects the infant's sleep data and evaluates the sleep quality. When the sleep quality is lower than expected, the trained environmental automatic control model is used to adjust the key factors affecting the infant's sleep quality, and a sleep health report is constructed based on the infant's sleep data. The abnormality detection unit detects abnormalities in the collected infant status data, activates a multi-level alarm mechanism and issues a multi-level alarm command when an abnormal event occurs, and switches the current alarm channel if the alarm command is not responded to in time; Abnormal alarm unit, if there are multiple abnormal events in the current connection, the alarm priority of each abnormal event is determined according to the degree of abnormality, and each guardian is selected and switched when issuing an alarm command according to the guardian's location data and response data; The feedback control unit sends the alarm instruction and the corresponding emergency plan to the selected guardian for confirmation. According to the baby's status condition data and alarm feedback data, the alarm sensitivity and the guardian who receives the alarm instruction are adjusted. The Analytic Hierarchy Process (AHP) is a decision-making method that treats complex multi-objective decision-making problems as a system, decomposes them into multiple levels of objectives, criteria or plans, and conducts qualitative and quantitative analysis on this basis. This method was proposed by Professor Satie, an American operations researcher in the early 1970s. It decomposes the decision-making problem into different hierarchical structures in the order of overall goal, sub-goals at each level, evaluation criteria, and specific alternative plans, and then solves the eigenvector of the judgment matrix to obtain the priority weight of each element at each level to an element at the previous level. Finally, the final weight of each alternative plan to the overall goal is recursively merged by the weighted sum method to select the best plan. The analytic hierarchy process is particularly effective in decision-making in social, economic and management fields.

[0042] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0043] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0044] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0045] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0046] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for intelligent monitoring of infants, characterized in that: include, The sensor network collects environmental data of the monitoring area, and constructs environmental coefficients based on continuous environmental data. , if the environmental factor If the environmental threshold is lower than the threshold, a monitoring instruction will be issued to the outside world; After collecting the baby's sleep data, the sleep quality is evaluated. When the sleep quality is lower than expected, the trained automatic environmental control model is used to adjust the key factors affecting the baby's sleep quality, and a sleep health report is constructed based on the baby's sleep data; Perform abnormal detection on the collected infant status data, activate the multi-level alarm mechanism and issue multi-level alarm instructions when abnormal events occur, and switch the current alarm channel if the alarm instruction is not responded to in time; If there are multiple abnormal events, the alarm priority of each abnormal event is determined according to the degree of abnormality, and each guardian is selected and switched when issuing an alarm command based on the guardian's location data and response data; The alarm instruction and the corresponding emergency plan are sent to the selected guardian and confirmed by the guardian. According to the infant status condition data and the alarm feedback data, the alarm sensitivity and the guardian who receives the alarm instruction are adjusted.

2. The method for intelligent monitoring of infants according to claim 1, characterized in that: The sensor network arranged in the monitoring area collects environmental data in the monitoring area and aggregates them to generate an environmental data set. The environmental data in the monitoring area is used as input, and the trained environmental evaluation model is used to evaluate the environmental conditions in the monitoring area and obtain the corresponding environmental score.

3. The method for intelligent monitoring of infants according to claim 2, characterized in that: The monitoring device located in the monitoring area records the sleep quality data of the infant in the monitoring area, and uses the sleep quality data as input to evaluate the sleep quality of the infant using the trained quality evaluation model to obtain a sleep quality value; Principal component analysis is used to screen out the key factors that affect the infant's sleep quality from various environmental parameters in the monitoring area, and the corresponding degree of influence is determined; when the infant's sleep quality is lower than expected, the trained environmental automatic control model is used to prioritize the adjustment and optimization of the key factors.

4. The method for intelligent monitoring of infants according to claim 3, characterized in that: After a preset observation period, a sleep health report is constructed based on the baby's sleep data, and the reporting interval of the sleep health report is constrained , the constraints are as follows: ； in, b is the number of sleep quality values, ; It is i The sleep quality value and j The difference between the sleep quality values, is the average of the differences, , ,and .

5. The method for intelligent monitoring of infants according to claim 4, characterized in that: The monitoring equipment located in the monitoring area collects the monitoring parameters of the infant, including sleep quality data, activity status data and vital sign data; the infant monitoring parameters are used as input and the trained anomaly detection model is used for detection. When an abnormal event is detected, a first-level alarm command is issued to the outside.

6. The method for intelligent monitoring of infants according to claim 5, characterized in that: After multiple first-level alarm instructions are issued continuously, under dimensionless conditions, the warning value is constructed based on the issuance status data of the first-level alarm instructions. , if the warning value obtained When the preset alarm threshold is exceeded, a secondary alarm command is issued to the outside; the alarm value is constructed based on the status data of the issuance of the primary alarm command. The way is as follows: ； Where: For the i Second and i +1 time interval between alarms, For the i The abnormality level at the time of the alarm, and are weight coefficients, , , is the attenuation coefficient of the cumulative impact, .

7. The method for intelligent monitoring of infants according to claim 6, characterized in that: After detecting that the infant's state is abnormal, the abnormal data of the abnormal event is used as input, and the trained abnormal evaluation model is used to perform abnormal evaluation and obtain the abnormal value; Each abnormal event is marked with an abnormal value, and after sorting each abnormal event according to the size of the abnormal value, the corresponding alarm priority is obtained.

8. The method for intelligent monitoring of infants according to claim 7, characterized in that: Obtain the location information of each guardian, obtain the distance data between the guardian and the baby, and the response speed of each guardian after receiving the alarm command; construct the responsiveness based on the distance data and response speed , according to the responsiveness Sort each guardian from small to large and obtain the corresponding alarm ranking.

9. The method for intelligent monitoring of infants according to claim 8, characterized in that: Distance and response speed Perform linear normalization and project the corresponding data values ​​onto The response is generated according to the following formula : ； Weight coefficient: , ,and , , The number of response speeds ， For the i The response speed, is the mean response speed, Its expected value.

10. The method for intelligent monitoring of infants according to claim 9, characterized in that: Formulate emergency response plans for each abnormal event. After an abnormal event occurs and an alarm instruction is issued to the outside, the alarm instruction and the corresponding emergency plan will be sent to the selected guardian for confirmation by the guardian; If the confirmation time exceeds the preset time, resend or switch the guardian, and collect the corresponding alarm feedback data. If the guardian fails to confirm twice after receiving the alarm command, he will no longer receive the alarm command.

11. The method for intelligent monitoring of infants according to claim 10, characterized in that: After a preset adjustment period, the warning threshold is adjusted according to the environmental status and sleep quality. Adjustments are made to improve the alarm sensitivity as the optimization goal, and the qualified standard values ​​of various monitoring parameters are optimized according to the alarm feedback data using a pre-trained genetic algorithm to obtain the optimized qualified standard values.

12. The method for intelligent monitoring of infants according to claim 11, characterized in that: Warning threshold The adjustment is as follows: ； Where: is the basic value of the warning threshold, In the time window Any time node, k is the sleep adjustment coefficient, For the time window Internal sleep quality value Below quality threshold The number of times is the rate of change of environmental conditions, is the rate of change adjustment factor.

13. An intelligent infant monitoring system, characterized in that: include, Environmental analysis unit, which collects environmental data of the monitoring area through the sensor network and constructs environmental coefficients based on continuous environmental data , if the environmental factor If the environmental threshold is lower than the threshold, a monitoring instruction will be issued to the outside world; The environmental condition adjustment unit collects the infant's sleep data and evaluates the sleep quality. When the sleep quality is lower than expected, the trained environmental automatic control model is used to adjust the key factors affecting the infant's sleep quality, and a sleep health report is constructed based on the infant's sleep data. The abnormality detection unit detects abnormalities in the collected infant status data, activates a multi-level alarm mechanism and issues a multi-level alarm command when an abnormal event occurs, and switches the current alarm channel if the alarm command is not responded to in time; Abnormal alarm unit, if there are multiple abnormal events in the current connection, the alarm priority of each abnormal event is determined according to the degree of abnormality, and each guardian is selected and switched when issuing an alarm command according to the guardian's location data and response data; The feedback control unit sends the alarm instruction and the corresponding emergency plan to the selected guardian for confirmation. According to the infant status condition data and the alarm feedback data, the alarm sensitivity and the guardian who receives the alarm instruction are adjusted.

Images