Home-based care service method and device, storage medium and computer equipment

Abstract

The invention discloses a home-based care service method and device, a storage medium and computer equipment, relates to the technical field of digital medical treatment, and mainly aims to accurately match care service for a care object, avoid rescue delay of the care object and improve care experience of the care object. Comprising the steps that multi-mode home-based care health monitoring data of a target home-based care object are collected in real time, and the multi-mode home-based care health monitoring data comprise at least one of voice data, vital sign data, behavior monitoring data and surrounding environment data; performing fusion processing on the multi-mode home-based care health monitoring data to obtain home-based care health monitoring fusion features, and inputting the home-based care health monitoring fusion features into a preset service prediction model for service prediction to obtain a care service type required to be provided for the target home-based care object; and providing the old-age care service for the target home-based old-age care object based on the old-age care service type.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital medical treatment, in particular to a home-based care service method and device, a storage medium and computer equipment. BACKGROUND

[0002] With the aggravation of population aging, home-based care has become the choice of more and more elderly people, so it is necessary to customize care services for home-based care personnel.

[0003] At present, the care service is usually realized by the way that the staff of the care service center regularly visits the home of the care object. However, if the care object suddenly gets ill, the next visit is needed to be discovered, which may miss the golden rescue time, resulting in the delay of the care service. Meanwhile, the staff can only determine the state of the care object by inquiry, which is difficult to accurately define the service needed by the care object at this time, thereby reducing the quality of the care service and the experience of the care object. SUMMARY

[0004] The present application provides a home-based care service method and device, a storage medium and computer equipment, which can accurately match the care service for the care object, avoid the rescue delay of the care object, and improve the care experience of the care object.

[0005] According to a first aspect of the present application, a home-based care service method is provided, comprising:

[0006] real-time collection of multi-modal home-based care health monitoring data of a target home-based care object, wherein the multi-modal home-based care health monitoring data comprises at least one of voice data, vital sign data, behavior monitoring data and surrounding environment data;

[0007] fusion processing of the multi-modal home-based care health monitoring data to obtain home-based care health monitoring fusion features, and inputting the home-based care health monitoring fusion features into a preset service prediction model for service prediction to obtain a care service type needed to be provided for the target home-based care object;

[0008] providing care service for the target home-based care object based on the care service type.

[0009] Optionally, the care service type comprises at least one of an emotional companion service type, a medical rescue service type and a safety patrol service type.

[0010] providing care service for the target home-based care object based on the care service type, comprising:

[0011] providing emotional companion service for the target home-based care object based on the emotional companion service type.

[0012] based on the medical rescue service type, providing a medical rescue service for the target home-based elderly care object;

[0013] based on the security patrol service type, providing a security patrol service for the target home-based elderly care object.

[0014] Optionally, based on the emotional companion service type, providing an emotional companion service for the target home-based elderly care object, comprising:

[0015] acquiring historical multi-modal home-based elderly care health monitoring data of the target home-based elderly care object in a normal state in a historical time period, and based on the historical multi-modal home-based elderly care health monitoring data, establishing an emotional baseline of the target home-based elderly care object;

[0016] based on the multi-modal home-based elderly care health monitoring data, determining an emotional monitoring line of the target home-based elderly care object, and determining a deviation degree between the emotional monitoring line and the emotional baseline;

[0017] based on the deviation degree, determining an emotional companion level, and dispatching an emotional companion strategy corresponding to the emotional companion level to provide an emotional companion service for the target home-based elderly care object.

[0018] Optionally, dispatching an emotional companion strategy corresponding to the emotional companion level to provide an emotional companion service for the target home-based elderly care object, comprising:

[0019] acquiring object feature data of the target home-based elderly care object, in the case that the emotional companion level is low, based on the object feature data, determining preference information of the target home-based elderly care object, and based on the preference information, sending robot voice greeting information to the target home-based elderly care object through an intelligent device, and pushing entertainment information;

[0020] in the case that the emotional companion level is medium, acquiring companion feature data of a plurality of to-be-assigned emotional companion personnel, and based on the companion feature data and the object feature data, determining a service effect score of each of the to-be-assigned emotional companion personnel corresponding to the target home-based elderly care object, based on the service effect score, performing emotional companion personnel assignment for the target home-based elderly care object, so as to perform communication and interaction between the assigned emotional companion personnel and the target home-based elderly care object;

[0021] In the case where the emotional companion level is high, the counseling personnel feature data of a plurality of to-be-assigned psychological counseling personnel is acquired, and based on the counseling personnel feature data and the object feature data, a psychological counseling effect score of each of the to-be-assigned psychological counseling personnel corresponding to the target home-based elderly care object is determined, the target home-based elderly care object is assigned a psychological counseling personnel based on the psychological counseling effect score, and the target home-based elderly care object is psychologically counseled by the assigned psychological counseling personnel.

[0022] Optionally, based on the medical rescue service type, medical rescue services are provided for the target home-based elderly care object, including:

[0023] Based on the multi-modal home-based elderly care health monitoring data, a health risk level of the target home-based elderly care object is determined, wherein the health risk level includes one of low risk, medium risk, and high risk.

[0024] A medical rescue strategy corresponding to the health risk level is dispatched to provide medical rescue services for the target home-based elderly care object, wherein the method of dispatching a medical rescue strategy corresponding to the health risk level to provide medical rescue services for the target home-based elderly care object includes:

[0025] In the case where the health risk level is low, health warning information and risk disposal suggestion information are generated based on the multi-modal home-based elderly care health monitoring data, and the health warning information and the risk disposal suggestion information are sent to a family terminal of the target home-based elderly care object, so that the family of the family terminal provides medical rescue services for the target home-based elderly care object based on the health warning information and the risk disposal suggestion information;

[0026] In the case where the health risk level is medium, first aid feature data of to-be-assigned community first aid personnel is acquired, and based on the multi-modal home-based elderly care health monitoring data and the first aid feature data, a first aid effect score of each of the to-be-assigned community first aid personnel corresponding to the target home-based elderly care object is determined, the target home-based elderly care object is assigned community first aid personnel based on the first aid effect score, and the assigned community first aid personnel is dispatched to provide medical rescue for the target home-based elderly care object.

[0027] In the case where the health risk level is high, historical medical record data of the target home-based elderly care object is acquired, intelligent outbound scripts are generated based on the multi-modal home-based elderly care health monitoring data and the historical medical record data, and the intelligent outbound scripts are used to intelligently call the medical emergency dispatch center, so that the target home-based elderly care object is sent to a hospital for medical rescue by the medical emergency dispatch center that passes the outbound call.

[0028] Optionally, the multi-modal home-based care health monitoring data is fused to obtain home-based care health monitoring fusion features, including:

[0029] The multi-modal home-based care health monitoring data corresponding home-based care health monitoring features and weight coefficients are determined respectively, the multi-modal home-based care health monitoring data corresponding home-based care health monitoring features are weighted and summed based on the weight coefficients, and the home-based care health monitoring fusion features are obtained.

[0030] Optionally, before the home-based care health monitoring fusion features are input into a preset service prediction model for service prediction, the method further includes:

[0031] An initial preset service prediction model is constructed;

[0032] A sample data set is obtained, wherein the sample data set includes multi-modal home-based care health monitoring data of sample home-based care objects with service type labels;

[0033] The sample data set is divided into a training set and a test set, the initial preset service prediction model is trained using the training set, and the trained initial preset service prediction model is tested using the test set, and finally the trained initial preset service prediction model that meets the test condition is used as the preset service prediction model.

[0034] According to a second aspect of the present application, a home-based care service device is provided, including:

[0035] A data acquisition unit is configured to acquire multi-modal home-based care health monitoring data of a target home-based care object in real time, wherein the multi-modal home-based care health monitoring data includes at least one of voice data, vital sign data, behavior monitoring data, and surrounding environment data;

[0036] A prediction unit is configured to fuse the multi-modal home-based care health monitoring data to obtain home-based care health monitoring fusion features, and input the home-based care health monitoring fusion features into a preset service prediction model for service prediction to obtain a care service type to be provided for the target home-based care object;

[0037] A care service unit is configured to provide care services for the target home-based care object based on the care service type.

[0038] According to a third aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the program is executed by a processor to implement the above home-based care service method.

[0039] According to a fourth aspect of the present application, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above home-based care service method when executing the program.

[0040] According to the home-based care service method, device, storage medium and computer device provided by the present application, compared with the current mode of realizing the care service by the staff of the care service center regularly checking the home of the care object, the present application realizes the care service by collecting the multi-modal home-based care health monitoring data of the target home-based care object in real time, analyzing and processing the multi-modal data by using a model to identify the care service type required by the target home-based care object, and finally providing the care service to the care object according to the care service type. Through the monitoring of the multi-modal health data, the health status and living conditions of the home-based care object can be comprehensively monitored. Through the real-time monitoring of the multi-modal health data, the dynamic changes of the home-based care object can be grasped in time, and once an abnormal situation occurs, corresponding measures can be taken immediately to avoid delaying the best treatment opportunity. Through the model, the care service type required by the care object can be predicted, which can ensure that the adaptability of the care service type to the care object is higher, and thus the care experience of the care object is improved. BRIEF DESCRIPTION OF DRAWINGS

[0041] The accompanying drawings, which are included to provide a further understanding of the present application and are incorporated in and constitute a part of this application, illustrate embodiments of the present application and serve to explain the principles of the present application, and do not limit the present application in any manner. In the drawings:

[0042] Figure 1 A flow chart of a home-based care service method provided by an embodiment of the present application is shown;

[0043] Figure 2 A flow chart of another home-based care service method provided by an embodiment of the present application is shown;

[0044] Figure 3 A structural schematic diagram of a home-based care service device provided by an embodiment of the present application is shown;

[0045] Figure 4 A structural schematic diagram of another home-based care service device provided by an embodiment of the present application is shown;

[0046] Figure 5 An entity structural schematic diagram of a computer device provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0047] Hereinafter, the present application will be described in detail with reference to the accompanying drawings and embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0048] At present, the mode of realizing the pension service is to have the staff of the pension service center regularly check the home of the pension object. There may be a situation that the abnormal situation is not found in time, resulting in the delay of the pension service. At the same time, due to the lack of experience and negligence of the staff, it is difficult to accurately define the service required by the pension object at this time, which reduces the quality of the pension service and the experience of the pension object.

[0049] In order to solve the above problems, the embodiment of the present application provides a home-based pension service method, as shown in the figure, the method comprises: Figure 1

[0050] 101, real-time collection of multi-modal home-based pension health monitoring data of a target home-based pension object, wherein the multi-modal home-based pension health monitoring data comprises at least one of voice data, vital sign data, behavior monitoring data and surrounding environment data.

[0051] Among them, the voice data refers to the call data between the target home-based pension object and the intelligent home-based pension device; the vital sign data refers to the heart rate, blood oxygen, body temperature, breathing frequency, body movement data and the like of the pension object; the behavior monitoring data refers to the data such as walking, falling, entering and exiting the door of the pension object; the surrounding environment data refers to the temperature, brightness, water leakage condition, smoke condition and the like of the area where the pension object is located.

[0052] ​This invention provides elderly care services through a smart speaker-based home care system. The hardware includes a smart speaker main unit (containing a microphone array and speaker), with an external device interface (supporting Bluetooth / Wi-Fi connection to smartwatches, vital sign monitors, etc.). On the software side, the elderly care service system employs a Transformer-based NLP model, supporting dialect recognition and contextual understanding. It uses an LSTM neural network-based respiratory rate detection algorithm for respiratory monitoring and provides an emotional communication platform for the elderly through digital human-based voice chat technology, alleviating feelings of loneliness. The elderly care service system also includes: a voice interaction module: using advanced voice recognition and natural language processing technology to enable voice interaction with the elderly, supporting functions such as voice control of home appliances, information retrieval, and sending emergency requests; a health monitoring module: integrating smartwatches, vital sign monitors, fall detectors, smoke alarms, and abnormal behavior detectors to monitor the health status and home safety of the elderly in real time, automatically triggering alarms when data is abnormal; an emotional companionship module: employing digital human voice chat technology, combined with real-person voice and video chat services, to provide an emotional exchange platform for the elderly and alleviate loneliness; and a disease encyclopedia module: containing a comprehensive database of common diseases among the elderly, supporting voice queries to help the elderly and their families understand disease prevention. Treatment knowledge; Service management module: Provides dedicated butlers to serve customers one-on-one, offering personalized life services and application usage guidance, and links with community hospitals, housekeeping platforms, and emergency rescue systems to achieve health warnings, service appointments, and remote medical guidance; Context understanding unit: Analyzes user intent through dialogue history and supports correction of fuzzy commands; Multimodal interaction unit: Combines voice and LED screen display to provide visualized health data feedback; Non-contact vital sign monitoring: Based on the principle of electromagnetic wave reflection, it monitors vital signs by emitting millimeter-wave signals and receiving signals reflected back from the human body; Fall detection algorithm: Identifies abnormal movements and other behavioral data and surrounding environmental data through environmental sensors (accelerometers, gyroscopes), and includes alarm functions.

[0053] This invention, through monitoring multimodal home-based elderly care health data, can comprehensively monitor the health status and living conditions of home-based elderly care recipients, enabling the subsequent determination of elderly care service types to better meet the needs of the elderly. Furthermore, real-time monitoring data can promptly grasp the dynamic changes in the condition of home-based elderly care recipients, and any abnormalities can be detected immediately and corresponding measures can be taken to avoid delaying the best treatment opportunity.

[0054] 102. The multimodal home-based elderly care health monitoring data is fused and processed to obtain the fused features of home-based elderly care health monitoring. The fused features of home-based elderly care health monitoring are then input into the preset service prediction model to predict the types of elderly care services that need to be provided to the target home-based elderly care recipients.

[0055] For the embodiments of the present application, in order to improve the prediction accuracy of the preset service prediction model, it is necessary to first train and construct the preset service prediction model. Based on this, the method comprises: constructing a preset initial service prediction model; obtaining a sample data set, wherein the sample data set comprises multi-modal home health monitoring data of sample home elderly objects with service type labels; dividing the sample data set into a training set and a test set, training the preset initial service prediction model using the training set, and testing the trained preset initial service prediction model using the test set, and finally taking the trained preset initial service prediction model that meets the test conditions as the preset service prediction model.

[0056] Specifically, in the model training process, first, a preset initial service prediction model is constructed, and then a sample data set is obtained. Ensure that the data set contains all necessary files, including multi-modal home health monitoring data of multiple sample elderly objects and their corresponding elderly service labels. Convert the data into a format that the preset initial service prediction model can understand, and finally train and test the model. Specifically, the data set can be divided first: use random or specific strategies (such as stratified sampling) to divide the sample data set into a training set and a test set. Then use the training set to train the model, and use the test set to test the trained model to evaluate its performance on unseen data. Calculate and record the mCP, accuracy, recall rate and other indicators on the test set. If the model performance does not meet the requirements, it can return to the training stage for more iterations or adjustments. The preset service prediction model that meets the requirements is obtained in this way. Further, the multi-modal home health monitoring data needs to be fused and processed, based on which step 102 specifically comprises: determining the home health monitoring features and weight coefficients corresponding to the multi-modal home health monitoring data respectively, weighting and summing the home health monitoring features corresponding to the multi-modal home health monitoring data based on the weight coefficients, and obtaining the home health monitoring fusion features.

[0057] Specifically, the home-based health monitoring data of each modality is converted into text data, and then a word embedding or CNN feature extraction model is used to extract the home-based health monitoring features of the home-based health monitoring data of each modality. According to the actual needs, weight coefficients are set for the home-based health monitoring data of different modalities. Finally, based on the weight coefficients, the home-based health monitoring features of each modality are weighted and summed to obtain the home-based health monitoring fusion features. Finally, the home-based health monitoring fusion features are directly input into a preset service prediction model, and the type of service required by the target home-based elderly object can be directly output by the preset service prediction model. The embodiment of the present application can fully utilize the relationship between data, extract more implicit features, make more sufficient use of data, and obtain more accurate prediction results. At the same time, the model is used to predict the service type, which can improve the prediction efficiency of the service type and make the predicted service type more suitable for the elderly object, thereby improving the experience of the elderly object.

[0058] 103. Based on the type of service for the elderly, providing service for the target home-based elderly object.

[0059] For the embodiment of the present application, after the model predicts the type of service required by the elderly object, the elderly object is provided with the service according to the type of service. For example, after the fall detection is triggered, the system automatically contacts the preset guardian and the 120 emergency center. The health warning mechanism: when abnormal heart rate is detected for three times in succession, the warning is pushed to the community hospital platform and the emergency family contact person is contacted. The embodiment of the present application can monitor the health status and living conditions of the home-based elderly object through the monitoring of multi-modal health data. Through real-time monitoring of multi-modal health data, the dynamic changes of the home-based elderly object can be grasped in time. Once an abnormal situation occurs, it can be detected immediately and appropriate measures can be taken to avoid delaying the best treatment opportunity. The model is used to predict the type of service required by the elderly object, which can ensure that the type of service is more suitable for the elderly object, thereby improving the experience of the elderly object.

[0060] According to the home-based care service method provided by the application, compared with the current mode of realizing the care service by the staff of the care service center regularly checking the home of the care object, the application realizes the care service by collecting the multi-modal home-based care health monitoring data of the target home-based care object in real time, and using a model to analyze and process the multi-modal data to identify the care service type required by the target home-based care object, and finally providing the care service for the care object according to the care service type. Through the monitoring of the multi-modal health data, the health status and living conditions of the home-based care object can be comprehensively monitored. Through the real-time monitoring of the multi-modal health data, the dynamic changes of the home-based care object can be grasped in time, and once an abnormal situation occurs, the corresponding measures can be taken immediately to avoid delaying the best treatment opportunity. Through the model, the care service type required by the care object can be predicted, which can ensure that the adaptability of the care service type to the care object is higher, and thus the care experience of the care object is improved.

[0061] Further, in order to better illustrate the above process of providing service for home-based care, as a refinement and expansion of the above embodiment, the embodiment of the application provides another home-based care service method, as shown in Figure 2 The method comprises the following steps:

[0062] 201, collecting multi-modal home-based care health monitoring data of a target home-based care object in real time, wherein the multi-modal home-based care health monitoring data comprises at least one of voice data, vital sign data, behavior monitoring data and surrounding environment data.

[0063] Specifically, the real-time monitoring and collection of the multi-modal home-based care health data are realized by the home-based care service system based on the intelligent sound box.

[0064] 202, performing fusion processing on the multi-modal home-based care health monitoring data to obtain home-based care health monitoring fusion features, and inputting the home-based care health monitoring fusion features into a preset service prediction model to perform service prediction, so as to obtain the care service type required to be provided for the target home-based care object.

[0065] For the embodiment of the present application, in order to carry out the prediction of the type of pension service, the multi-modal home pension health monitoring data needs to be fused and processed, based on which, step 202 specifically includes: if the multi-modal home pension health monitoring data includes voice data, vital sign data and behavior monitoring data, the voice data corresponding voice feature vector, the vital sign data corresponding vital sign feature vector and the behavior monitoring data corresponding behavior feature vector are determined respectively; the voice feature vector, the vital sign feature vector and the behavior feature vector are subjected to feature-level fusion processing to obtain a feature fusion vector; the voice feature vector, the vital sign feature vector and the behavior feature vector are subjected to element-level fusion processing to obtain an element fusion vector; the voice feature vector, the vital sign feature vector and the behavior feature vector are subjected to low-order fusion processing to obtain a low-order fusion vector; and the feature fusion vector, the element fusion vector and the low-order fusion vector are combined to obtain a home pension health monitoring fusion feature.

[0066] Specifically, the speech data corresponding speech feature vector, the vital sign data corresponding vital sign feature vector, the behavior monitoring data corresponding behavior feature vector, and the surrounding environment data corresponding environment feature vector are determined by using the CNN feature extraction and the like. Then, in order to make full use of the relationship between the data, more implicit features are extracted, and the high-order and low-order processing is taken into account, so that the data is more fully utilized, and the prediction result obtained later is more accurate, meeting the needs of the actual application scene. The speech feature vector, the vital sign feature vector, the behavior feature vector, and the environment feature vector need to be cross-processed. The specific cross-processing method includes: cross-processing between different feature vectors at the feature level, that is, the Hadamard product of all elements between the vectors is performed, and then convolution transformation is performed under a certain weight to obtain a feature cross vector f(w*(a1*b1*c1, a2*b2*c3, a3*b3*c3)); at the same time, the element level fusion of all feature vectors is performed, that is, the Hadamard product of each element between the vectors is performed, and then different weight values are assigned to the results after each product, and linear transformation is performed to obtain an element cross vector f(w1*a1*b1*c1, w2*a2*b2*c2, w3*a3*b3*c3); in addition, all feature vectors are subjected to low-order fusion processing, and a weight coefficient is assigned to the result after cross-processing, and linear transformation is performed to obtain a low-order cross vector f(w4(a1, a2, b1, b2, c1, c2)); finally, the above feature fusion vectors, element fusion vectors, and low-order fusion vectors are combined together, such as transverse splicing, to obtain a home-based elderly health monitoring fusion feature. It should be noted that the above examples are only illustrative and do not limit the embodiments of the present application. Thus, by cross-processing the speech feature vector, the vital sign feature vector, and the behavior feature vector, different features can be automatically or explicitly combined to generate new feature combinations. These combined features can contain complex nonlinear relationships between original features, so that the model can capture more detailed and rich information in the data, that is, the relationship between various data can be fully utilized to extract more implicit features, while the high-order and low-order processing is taken into account, so that the data is more fully utilized, and the service type prediction result obtained later is more accurate, meeting the needs of the actual application scene. Finally, the preset service prediction model is directly used to analyze the home-based elderly health monitoring fusion feature to determine the suitable service type for the elderly object.

[0067] 203、if the service type is an emotional companion service type, an emotional companion service is provided for the target home-based elderly object based on the emotional companion service type.

[0068] The emotional accompanying service refers to emotional communication such as chatting and psychological counseling with the elderly. For the embodiment of the present application, if the model predicts that the current elderly needs emotional accompanying service, the target home-based elderly needs to provide emotional accompanying service. Based on this, step 203 specifically includes: obtaining historical multi-modal home-based elderly health monitoring data of the target home-based elderly in a historical period under a normal state, and establishing an emotional baseline of the target home-based elderly based on the historical multi-modal home-based elderly health monitoring data; determining an emotional monitoring line of the target home-based elderly based on the multi-modal home-based elderly health monitoring data, and determining a deviation degree between the emotional monitoring line and the emotional baseline; determining an emotional accompanying level based on the deviation degree, and dispatching an emotional accompanying strategy corresponding to the emotional accompanying level to provide emotional accompanying service for the target home-based elderly.

[0069] The normal state refers to a state in which the elderly has no abnormal behavior and is not lonely.

[0070] Specifically, a continuous time period (e.g., January- June 2023) in which the target home-based elderly person has not reported emotional abnormalities in the past 6 months is selected, voice data: 30 minutes of conversation is collected daily through a smart speaker (sampling rate 16 kHz, 16 bit precision), acoustic features (fundamental frequency, speech rate, energy, MFCC coefficients) and other data are extracted; behavior data: the number of times of turning over at night (≤2 times per hour is normal), the length of daytime activity (≥2 hours per day is normal) and other data are monitored by millimeter wave radar; environmental data: indoor temperature (20-26℃ is the comfortable interval), light intensity (300-500 lux is the appropriate reading interval) and other data are recorded by temperature and humidity sensors. The historical data segment is labeled, for example: label 1: calm (HR = 60-75 bpm, speech rate = 120-150 words / minute, turning over once at night); label 2: slight joy (HR = 70-85 bpm, speech rate = 150-180 words / minute, daytime activity for 3 hours); label 3: anxiety (HR = 90-110 bpm, speech rate = 180-220 words / minute, EDA fluctuation > 30%). The mean value of the statistical features of the historical data segment labeled as "calm" is calculated to form the baseline: for example, the voice baseline: mean value of fundamental frequency = 120 Hz, mean value of speech rate = 135 words / minute; the behavior baseline: the number of times of turning over at night = 1.2 times / hour, the length of daytime activity = 2.5 hours / day, the environmental baseline: indoor temperature = 23℃±1℃, light intensity = 400 lux±50 lux. The above various baselines can be collectively referred to as emotional baselines, and based on the collected multi-modal home-based elderly health monitoring data, the emotional monitoring line of the elderly person can be determined, for example, the voice monitoring line: current speech rate = 190 words / minute, fundamental frequency = 145 Hz, the behavior monitoring line: turning over at night = 3 times / hour (current time is 2 am); the environmental monitoring line: temperature = 28℃, light = 150 lux (abnormal bright light at night). Compare each monitoring line with the corresponding baseline, and determine the emotional state of the elderly person according to the degree of deviation between the two, and then determine the emotional care level according to the emotional state, for example, if the deviation between the monitoring line and the corresponding baseline is large, it is determined that the elderly person is currently in a depressed state, and the emotional care level can be set to high level, if the deviation between the monitoring line and the corresponding baseline is moderate, it is determined that the elderly person is currently in an anxious state, and the emotional care level can be set to middle level, if the deviation between the monitoring line and the corresponding baseline is small, it is determined that the elderly person is currently relatively calm but very lonely, and the emotional care level can be set to low level.

[0071] Further, after determining the emotional companion level, it is also necessary to dispatch an emotional companion strategy corresponding to the emotional companion level to provide emotional companion services for the target home-based elderly object. Based on this, the method comprises: acquiring object characteristic data of the target home-based elderly object; in the case that the emotional companion level is low, determining the preference information of the target home-based elderly object based on the object characteristic data, and based on the preference information, sending robot voice greeting information and entertainment information to the target home-based elderly object through an intelligent device; in the case that the emotional companion level is medium, acquiring companion worker characteristic data of a plurality of to-be-assigned emotional companion workers, and based on the companion worker characteristic data and the object characteristic data, determining a service effect score of each of the to-be-assigned emotional companion workers corresponding to the target home-based elderly object, performing emotional companion worker assignment for the target home-based elderly object based on the service effect score, so as to perform communication and interaction between the assigned emotional companion worker and the target home-based elderly object; in the case that the emotional companion level is high, acquiring counseling worker characteristic data of a plurality of to-be-assigned psychological counseling workers, and based on the counseling worker characteristic data and the object characteristic data, determining a psychological counseling effect score of each of the to-be-assigned psychological counseling workers corresponding to the target home-based elderly object, performing psychological counseling worker assignment for the target home-based elderly object based on the psychological counseling effect score, so as to perform psychological counseling on the target home-based elderly object through the assigned psychological counseling worker.

[0072] Among them, the object characteristic data comprises age, gender, interest and hobby of the elderly object; the preference information refers to the entertainment type and greeting sentence preferred by the elderly object; the companion worker characteristic data comprises but is not limited to age, gender, previous companion object and companion quality of the companion worker; the counseling worker characteristic data comprises but is not limited to age, gender, psychological counseling experience and psychological counseling quality, and user feedback information of the counseling worker.

[0073] Specifically, in the case of a low emotional companion level, a voice greeting, entertainment information liked by the object, etc. are played to the user through a smart speaker device, etc. In the case of a medium emotional companion level, an emotional companion needs to be dispatched to provide emotional companionship to the elderly object. In order to improve the quality of emotional companionship, it is necessary to reasonably select an emotional companion. Based on this, the method comprises: determining an object feature vector corresponding to the object feature data and a companion feature vector corresponding to the companion feature data, and respectively performing fusion processing on the object feature vector and each of the companion feature vectors to obtain a fusion feature vector between the target elderly object and each of the to-be-assigned emotional companions; each of the fusion feature vectors is input into a preset service effect score model for score prediction to obtain a service effect score of each of the to-be-assigned emotional companions corresponding to the target elderly object. Finally, the emotional companion corresponding to the highest service effect score is selected to provide emotional companionship to the elderly object, and the mode of emotional companionship can be voice or video chat. In the case of a high emotional companion level, a psychological counselor needs to be dispatched to provide psychological counseling to the elderly object. In order to improve the quality of psychological counseling, it is necessary to reasonably select a psychological counselor. Based on this, the method comprises: determining an object feature vector corresponding to the object feature data and a counseling feature vector corresponding to the counseling feature data, and respectively performing fusion processing on the object feature vector and each of the counseling feature vectors to obtain a fusion feature vector between the target elderly object and each of the to-be-assigned psychological counselors; each of the fusion feature vectors is input into a preset counseling effect score model for score prediction to obtain a counseling effect score of each of the to-be-assigned psychological counselors corresponding to the target elderly object. Finally, the psychological counselor corresponding to the highest counseling effect score is selected to provide psychological counseling to the elderly object. The embodiments of the present application set the emotional companion level for the elderly object, and different emotional companion levels adopt different emotional companion strategies, which can ensure the emotional companion effect and avoid resource waste. It should be noted that the preset service effect score model and the preset counseling effect score model in the embodiments of the present application are constructed in advance based on a sample data set. For example, the sample data set of the preset service effect score model includes feature data of emotional companions and feature data of elderly objects, and actual service effect scores of each emotional companion for the elderly object. The sample data set of the preset counseling effect score model includes feature data of psychological counselors and feature data of elderly objects, and actual counseling effect scores of each psychological counselor for the elderly object.

[0074] 204、If the elderly service type is a medical rescue service type, a medical rescue service is provided for the target home-based elderly object based on the medical rescue service type.

[0075] For the embodiment of the application, when the pension service type is a medical rescue service type, medical rescue needs to be performed on the pension object. Based on this, step 204 specifically includes: determining a health risk level of the target home-based pension object based on the multi-modal home-based pension health monitoring data, wherein the health risk level includes one of low risk, medium risk, and high risk; and dispatching a medical rescue strategy corresponding to the health risk level to provide medical rescue services for the target home-based pension object. The method of dispatching a medical rescue strategy corresponding to the health risk level to provide medical rescue services for the target home-based pension object includes: in the case of a low risk level, generating health warning information and risk disposal suggestion information based on the multi-modal home-based pension health monitoring data, and sending the health warning information and the risk disposal suggestion information to a family terminal of the target home-based pension object, so that the family of the family terminal performs medical rescue services on the target home-based pension object based on the health warning information and the risk disposal suggestion information; in the case of a medium risk level, obtaining first aid characteristic data of a to-be-assigned community first aid worker, and determining an emergency effect score of each to-be-assigned community first aid worker corresponding to the target home-based pension object based on the multi-modal home-based pension health monitoring data and the first aid characteristic data, performing community first aid worker assignment based on the emergency effect score, and dispatching the assigned community first aid worker to perform medical rescue on the target home-based pension object; and in the case of a high risk level, obtaining historical medical record data of the target home-based pension object, generating intelligent outbound call scripts based on the multi-modal home-based pension health monitoring data and the historical medical record data, and performing intelligent outbound call on a medical emergency dispatch center based on the intelligent outbound call scripts, so that the medical emergency dispatch center that passes the outbound call sends the target home-based pension object to a hospital for medical rescue.

[0076] The emergency characteristic data includes professional skills (such as cardiopulmonary resuscitation, trauma first aid, etc.), work experience, current location, and deployable time of the emergency personnel. Specifically, for example, if the elderly object has only fallen down and has not fainted or shown other abnormalities, it can be determined that the health risk level includes low risk, at which time health warning information for falling down and disposal suggestion information (such as preparing band-aids and other medicines) for falling down are generated, and the above information is sent to the family members of the elderly object through short message, telephone, etc., so as to let the family members handle the abnormal situation. If the elderly object has a persistent rapid or slow heart rate and exceeds the normal range by a large margin, it is determined that the health risk level of the elderly object is at a medium risk level. Based on the multi-modal home-based elderly health monitoring data and the emergency characteristic data of the community emergency personnel to be allocated, a certain scoring algorithm is used to calculate the emergency effect score of each community emergency personnel corresponding to the target home-based elderly object. The scoring algorithm can consider multiple factors, for example, if the main problem of the target home-based elderly object is heart-related abnormalities, then the emergency personnel with cardiopulmonary resuscitation professional skills and rich experience will obtain a higher score; at the same time, the distance between the emergency personnel and the target home-based elderly object is considered, and the closer the distance, the higher the score, so as to ensure that the emergency personnel can quickly arrive at the scene to perform rescue. According to the calculated emergency effect score, the community emergency personnel with the highest score is selected to be allocated to the target home-based elderly object. The system sends rescue task information to the terminal of the allocated community emergency personnel through a wireless communication network, including the address of the target home-based elderly object, a summary of the health condition, an estimated arrival time, etc. After receiving the task, the community emergency personnel quickly goes to the residence of the target home-based elderly object to perform medical rescue. If the elderly object has syncope, it is determined that the health risk level of the elderly object is at a high risk level. The system obtains historical medical record data from the electronic health record of the target home-based elderly object. The historical medical record data includes information such as past disease diagnosis, treatment record, allergy history, medication history, etc. Based on the multi-modal home-based elderly health monitoring data and the historical medical record data, a natural language processing technology is used to generate intelligent outbound call scripts. The intelligent outbound call script should include the basic information of the target home-based elderly object, the current health condition, the description of the emergency situation, and the measures that need to be taken by the medical emergency dispatch center, etc. For example, “Hello, this is the intelligent rescue system of [system name]. There is a home-based elderly object [old person's name], age [X] years old, with [past disease diagnosis] disease history. The current monitoring shows that [specific abnormal physiological indicators or symptoms], the situation is urgent, please arrange an ambulance to go to [detailed address] for rescue as soon as possible, and notify the nearby hospital to be prepared for admission.” The system converts the generated intelligent outbound call script into a voice signal through a voice synthesis technology, and uses an automatic outbound call system to dial the phone number of the medical emergency dispatch center. During the call, the system monitors the outbound state in real time to ensure that the medical emergency dispatch center can accurately receive the rescue information. At the same time, the system records the outbound time, call content, etc. for subsequent query and tracking.After the medical emergency dispatch center receives the outgoing call, the medical emergency dispatch center quickly arranges an ambulance and medical staff to go to the target home-based elderly person's residence according to the provided information, and sends the target home-based elderly person to a hospital for further medical rescue. The embodiment of the present application can accurately match the medical rescue mode for the elderly by selecting different medical rescue modes according to different health levels, so as to avoid waste of medical resources and ensure timely and effective treatment of the elderly.

[0077] 205、If the pension service type is a safety patrol service type, a safety patrol service is provided for the target home-based elderly person based on the safety patrol service type.

[0078] Specifically, a professional safety patrol team is formed according to the personnel configuration of the pension service institution and the number and distribution of the target home-based elderly person. The team members should include personnel with basic pension care knowledge, safety emergency handling ability and good communication skills. For example, a certain proportion of medical staff can be equipped in the team to perform preliminary treatment in case of sudden health problems during the patrol; at the same time, personnel familiar with safety assessment of home environment are arranged to be responsible for safety inspection of the living environment.

[0079] Compared with the current way of realizing pension service by having the staff of the pension service center regularly visit the homes of the elderly, the home-based pension service method provided by the present application can realize real-time collection of multi-modal home-based pension health monitoring data of the target home-based elderly person, and use a model to analyze and process the multi-modal data to identify the pension service type required by the target home-based elderly person, and finally provide pension service for the elderly according to the pension service type. Through the monitoring of multi-modal health data, the health status and living conditions of the home-based elderly person can be comprehensively monitored; through real-time monitoring of multi-modal health data, the dynamic changes of the home-based elderly person can be timely grasped, and once an abnormal situation occurs, appropriate measures can be taken immediately to avoid delay of the best treatment opportunity; through the model, the pension service type required by the elderly can be predicted, which can ensure that the pension service type is more suitable for the elderly, thereby improving the pension experience of the elderly.

[0080] Further, as a specific implementation of Figure 1 , the embodiment of the present application provides a home-based pension service device, as shown in Figure 3 , the device comprises a data acquisition unit 31, a prediction unit 32 and a pension service unit 33.

[0081] The data acquisition unit 31 can be used for real-time acquisition of multi-modal home-based pension health monitoring data of the target home-based elderly person, wherein the multi-modal home-based pension health monitoring data comprises at least one of voice data, vital sign data, behavior monitoring data and surrounding environment data.

[0082] The prediction unit 32 can be configured to perform fusion processing on the multi-modal home care health monitoring data to obtain home care health monitoring fusion features, and input the home care health monitoring fusion features into a preset service prediction model to perform service prediction, so as to obtain a type of care service to be provided for the target home care object.

[0083] The care service unit 33 can be configured to provide care service for the target home care object based on the type of care service.

[0084] In a specific application scenario, the type of care service includes at least one of an emotional companion service type, a medical rescue service type, and a safety patrol service type. In order to provide care service for the target home care object, the care service unit 33 can be configured to obtain historical multi-modal home care health monitoring data of the target home care object in a historical time period in a normal state, and establish an emotional baseline of the target home care object based on the historical multi-modal home care health monitoring data. Figure 4 As shown in the figure, the care service unit 33 includes an emotional companion module 331, a medical rescue module 332, and a safety patrol module 333.

[0085] The emotional companion module 331 can be configured to provide emotional companion service for the target home care object based on the emotional companion service type.

[0086] The medical rescue module 332 can be configured to provide medical rescue service for the target home care object based on the medical rescue service type.

[0087] The safety patrol module 333 can be configured to provide safety patrol service for the target home care object based on the safety patrol service type.

[0088] In a specific application scenario, in order to provide emotional companion service for the target home care object, the emotional companion module 331 can be specifically configured to obtain historical multi-modal home care health monitoring data of the target home care object in a historical time period in a normal state, and establish an emotional baseline of the target home care object based on the historical multi-modal home care health monitoring data. The emotional companion module 331 can be configured to determine an emotional monitoring line of the target home care object based on the multi-modal home care health monitoring data, and determine a deviation degree between the emotional monitoring line and the emotional baseline. The emotional companion module 331 can be configured to determine an emotional companion level based on the deviation degree, and dispatch an emotional companion strategy corresponding to the emotional companion level to provide emotional companion service for the target home care object.

[0089] In a specific application scenario, in order to take corresponding emotional accompanying strategies to provide emotional accompanying services for the target home-based elderly object, the emotional accompanying module 331 can be specifically used to obtain object characteristic data of the target home-based elderly object, determine preference information of the target home-based elderly object based on the object characteristic data in the case that the emotional accompanying level is low, and send robot voice greeting information to the target home-based elderly object through an intelligent device based on the preference information and push entertainment information; in the case that the emotional accompanying level is medium, obtain accompanying personnel characteristic data of a plurality of to-be-assigned emotional accompanying personnel, and determine a service effect score of each to-be-assigned emotional accompanying personnel corresponding to the target home-based elderly object based on the accompanying personnel characteristic data and the object characteristic data, perform emotional accompanying personnel assignment on the target home-based elderly object based on the service effect score, so as to perform communication and interaction with the target home-based elderly object through the assigned emotional accompanying personnel; in the case that the emotional accompanying level is high, obtain counseling personnel characteristic data of a plurality of to-be-assigned psychological counseling personnel, and determine a psychological counseling effect score of each to-be-assigned psychological counseling personnel corresponding to the target home-based elderly object based on the counseling personnel characteristic data and the object characteristic data, perform psychological counseling personnel assignment on the target home-based elderly object based on the psychological counseling effect score, so as to perform psychological counseling on the target home-based elderly object through the assigned psychological counseling personnel.

[0090] In a specific application scenario, in order to provide medical rescue services for the target home-based elderly object, the medical rescue module 332 can be specifically used to determine the health risk level of the target home-based elderly object based on the multi-modal home-based elderly health monitoring data, wherein the health risk level includes one of low risk, medium risk and high risk; dispatch the medical rescue strategy corresponding to the health risk level to provide medical rescue services for the target home-based elderly object, wherein the method of dispatching the medical rescue strategy corresponding to the health risk level to provide medical rescue services for the target home-based elderly object includes: in the case that the health risk level is low risk, generating health warning information and risk disposal suggestion information based on the multi-modal home-based elderly health monitoring data, and sending the health warning information and the risk disposal suggestion information to the family terminal of the target home-based elderly object, so that the family of the family terminal can provide medical rescue services for the target home-based elderly object based on the health warning information and the risk disposal suggestion information; in the case that the health risk level is medium risk, obtaining emergency characteristic data of a to-be-assigned community first-aid worker, and determining an emergency effect score of each to-be-assigned community first-aid worker corresponding to the target home-based elderly object based on the multi-modal home-based elderly health monitoring data and the emergency characteristic data, assigning the target home-based elderly object based on the emergency effect score, and dispatching the assigned community first-aid worker to provide medical rescue for the target home-based elderly object; in the case that the health risk level is high risk, obtaining historical medical record data of the target home-based elderly object, generating intelligent outbound scripts based on the multi-modal home-based elderly health monitoring data and the historical medical record data, and intelligently calling the medical emergency dispatch center based on the intelligent outbound scripts, so as to send the target home-based elderly object to a hospital for medical rescue through the medical emergency dispatch center passed by the outbound call.

[0091] In a specific application scenario, in order to fuse the multi-modal home-based elderly health monitoring data, the prediction unit 32 can be specifically used to determine the home-based elderly health monitoring features and weight coefficients corresponding to the multi-modal home-based elderly health monitoring data respectively, weight and sum the home-based elderly health monitoring features corresponding to the multi-modal home-based elderly health monitoring data based on the weight coefficients, and obtain the home-based elderly health monitoring fusion features.

[0092] In a specific application scenario, in order to construct a preset service prediction model, the device further includes a model construction unit 34.

[0093] The model construction unit 34 can be used to construct a preset initial service prediction model; obtain a sample data set, wherein the sample data set includes multi-modal home health monitoring data of sample home elderly people with service type labels; divide the sample data set into a training set and a test set, train the preset initial service prediction model by using the training set, test the trained preset initial service prediction model by using the test set, and finally take the trained preset initial service prediction model meeting the test condition as a preset service prediction model.

[0094] It should be noted that other corresponding descriptions of the functions of the home care service device provided by the embodiments of the present application can be referred to the corresponding descriptions of the method shown in Figure 1 The corresponding descriptions of the method shown in

[0095] Based on the above method as shown in Figure 1 Correspondingly, the embodiments of the present application also provide a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the following steps: real-time acquisition of multi-modal home health monitoring data of a target home elderly person, wherein the multi-modal home health monitoring data includes at least one of voice data, vital sign data, behavior monitoring data and surrounding environment data; fusion processing of the multi-modal home health monitoring data to obtain home health monitoring fusion features, and input of the home health monitoring fusion features into a preset service prediction model for service prediction to obtain a type of pension service needed to be provided for the target home elderly person; and provision of pension service for the target home elderly person based on the type of pension service.

[0096] Based on the above method as shown in Figure 1 and the device as shown in Figure 3 Based on the above method as shown in Figure 5As shown, the computer device comprises a processor 41, a memory 42, and a computer program stored on the memory 42 and executable on the processor, wherein the memory 42 and the processor 41 are both arranged on a bus 43, and the processor 41 implements the following steps when executing the program: collecting multi-modal home-based care health monitoring data of a target home-based care object in real time, wherein the multi-modal home-based care health monitoring data comprises at least one of voice data, vital sign data, behavior monitoring data, and surrounding environment data; performing fusion processing on the multi-modal home-based care health monitoring data to obtain home-based care health monitoring fusion features, and inputting the home-based care health monitoring fusion features into a preset service prediction model for service prediction to obtain a type of care service that needs to be provided for the target home-based care object; and providing care service for the target home-based care object based on the type of care service.

[0097] Through the technical scheme of the present application, the multi-modal home-based care health monitoring data of a target home-based care object is collected in real time, and a model is used to analyze and process the multi-modal data to identify the type of care service required by the target home-based care object, and finally the home-based care object is provided with care service according to the type of care service. Through the monitoring of multi-modal health data, the health status and living conditions of the home-based care object can be comprehensively monitored; through real-time monitoring of multi-modal health data, the dynamic changes of the home-based care object can be grasped in time, and once an abnormal situation occurs, it can be immediately detected and appropriate measures can be taken to avoid delaying the best treatment opportunity; through the model, the type of care service required by the home-based care object can be predicted, which can ensure that the adaptability of the type of care service to the home-based care object is higher, and thus the care experience of the home-based care object can be improved.

[0098] Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the present application can be realized by general computing devices, which can be concentrated on a single computing device or distributed on a network composed of multiple computing devices, and optionally, they can be realized by program codes executable by the computing devices, so that they can be stored in storage devices and executed by the computing devices, and in some cases, the steps shown or described can be executed in different orders, or they can be manufactured into individual integrated circuit modules, or multiple modules or steps thereof can be manufactured into a single integrated circuit module. Thus, the present application is not limited to any specific combination of hardware and software.

[0099] The above only describes the preferred embodiments of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A home nursing service method characterized by comprising: The method comprises: real-time collection of multi-modal home-based care health monitoring data of a target home-based care object, wherein the multi-modal home-based care health monitoring data comprises at least one of voice data, vital sign data, behavior monitoring data, and surrounding environment data; fusion processing of the multi-modal home-based care health monitoring data to obtain home-based care health monitoring fusion features, and inputting the home-based care health monitoring fusion features into a preset service prediction model for service prediction to obtain a type of care service to be provided for the target home-based care object; providing care service for the target home-based care object based on the type of care service.

2. The method of claim 1, wherein, The type of care service comprises at least one of an emotional companion service type, a medical rescue service type, and a safety patrol service type; providing care service for the target home-based care object based on the type of care service comprises: providing emotional companion service for the target home-based care object based on the emotional companion service type; providing medical rescue service for the target home-based care object based on the medical rescue service type; providing safety patrol service for the target home-based care object based on the safety patrol service type.

3. The method of claim 2, wherein, Providing emotional companion service for the target home-based care object based on the emotional companion service type comprises: acquiring historical multi-modal home-based care health monitoring data of the target home-based care object in a normal state in a historical time period, and establishing an emotional baseline of the target home-based care object based on the historical multi-modal home-based care health monitoring data; determining an emotional monitoring line of the target home-based care object based on the multi-modal home-based care health monitoring data, and determining a deviation degree between the emotional monitoring line and the emotional baseline; determining an emotional companion level based on the deviation degree, and dispatching an emotional companion strategy corresponding to the emotional companion level to provide emotional companion service for the target home-based care object.

4. The method of claim 3, wherein, Dispatching an emotional companion strategy corresponding to the emotional companion level to provide emotional companion service for the target home-based care object comprises: acquiring object feature data of the target home-based care object, determining preference information of the target home-based care object based on the object feature data in a case where the emotional companion level is low, and sending robot voice greeting information and pushing entertainment information to the target home-based care object through an intelligent device based on the preference information; in a case where the emotional companion level is medium, acquiring companion feature data of a plurality of to-be-assigned emotional companion personnel, determining a service effect score of each of the to-be-assigned emotional companion personnel corresponding to the target home-based care object based on the companion feature data and the object feature data, and performing emotional companion personnel assignment for the target home-based care object based on the service effect score to enable the assigned emotional companion personnel to communicate with the target home-based care object. In the case that the emotional companion level is high, the psychological counseling personnel feature data of multiple to-be-assigned psychological counseling personnel is acquired, and based on the psychological counseling personnel feature data and the object feature data, a psychological counseling effect score of each of the to-be-assigned psychological counseling personnel corresponding to the target home-based elderly care object is determined, the target home-based elderly care object is assigned with psychological counseling personnel based on the psychological counseling effect score, and the target home-based elderly care object is psychologically counseled by the assigned psychological counseling personnel.

5. The method of claim 2, wherein, Based on the medical rescue service type, medical rescue service is provided for the target home-based elderly care object, including: Based on the multi-modal home-based elderly care health monitoring data, the health risk level of the target home-based elderly care object is determined, wherein the health risk level includes one of low risk, medium risk and high risk; The medical rescue strategy corresponding to the health risk level is dispatched to provide medical rescue service for the target home-based elderly care object, wherein the method of dispatching the medical rescue strategy corresponding to the health risk level to provide medical rescue service for the target home-based elderly care object includes: In the case that the health risk level is low, health warning information and risk disposal suggestion information are generated based on the multi-modal home-based elderly care health monitoring data, and the health warning information and the risk disposal suggestion information are sent to the family terminal of the target home-based elderly care object, so that the family of the family terminal can provide medical rescue service for the target home-based elderly care object based on the health warning information and the risk disposal suggestion information; In the case that the health risk level is medium, the first aid feature data of the to-be-assigned community first aid personnel is acquired, and based on the multi-modal home-based elderly care health monitoring data and the first aid feature data, the first aid effect score of each of the to-be-assigned community first aid personnel corresponding to the target home-based elderly care object is determined, the target home-based elderly care object is assigned with community first aid personnel based on the first aid effect score, and the target home-based elderly care object is provided with medical rescue by the dispatched community first aid personnel; In the case that the health risk level is high, the historical medical record data of the target home-based elderly care object is acquired, intelligent outbound scripts are generated based on the multi-modal home-based elderly care health monitoring data and the historical medical record data, and the intelligent outbound scripts are used to intelligently call the medical emergency dispatch center, so that the target home-based elderly care object is sent to the hospital for medical rescue by the medical emergency dispatch center that passes the outbound call.

6. The method of claim 1, wherein, The multi-modal home-based elderly care health monitoring data is fused to obtain home-based elderly care health monitoring fusion features, including: The home-based elderly care health monitoring features corresponding to the multi-modal home-based elderly care health monitoring data and the weight coefficients are determined respectively, the home-based elderly care health monitoring features corresponding to the multi-modal home-based elderly care health monitoring data are weighted and summed based on the weight coefficients, and the home-based elderly care health monitoring fusion features are obtained.

7. The method of claim 1, wherein, Before the home-based elderly care health monitoring fusion features are input into a preset service prediction model for service prediction, the method further includes: constructing a preset initial service prediction model; obtaining a sample data set, wherein the sample data set comprises multi-modal home health monitoring data of sample home nursing objects with service type labels; dividing the sample data set into a training set and a test set, training the preset initial service prediction model using the training set, testing the trained preset initial service prediction model using the test set, and finally taking the trained preset initial service prediction model that meets the test condition as a preset service prediction model.

8. A home care service device characterized by comprising: comprise: a data acquisition unit configured to acquire multi-modal home health monitoring data of a target home nursing object in real time, wherein the multi-modal home health monitoring data comprises at least one of voice data, vital sign data, behavior monitoring data, and surrounding environment data; a prediction unit configured to perform fusion processing on the multi-modal home health monitoring data to obtain home health monitoring fusion features, input the home health monitoring fusion features into a preset service prediction model for service prediction, and obtain a type of pension service that needs to be provided for the target home nursing object; a pension service unit configured to provide pension services for the target home nursing object based on the type of pension service.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, implements the steps of the method of any one of claims 1 to 7.

10. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The computer program, when executed by a processor, implements the steps of the method of any one of claims 1 to 7.

Images