A medical monitoring-type elderly care robot

Abstract

This invention relates to the field of elderly care robot technology, specifically to a medical monitoring-type elderly care robot, comprising a physiological monitoring unit, an air particulate matter control and health feedback unit, an emotional and health integration unit, and a nursing assistance unit. The physiological monitoring unit collects the elderly person's physiological data in real time. The air particulate matter control unit detects pollutants and biological particles in the air, analyzes air quality in real time, and automatically adjusts the air purification equipment within the robot or the external ventilation device based on the concentration of airborne particles and the elderly person's physiological data. The emotional and health integration unit analyzes the elderly person's mental health. The nursing assistance unit includes a robotic arm and a voice control system to assist the elderly person in daily activities. This invention provides timely psychological support and emotional care for the elderly, improves their quality of life, and effectively prevents mental health problems.

Description

Technical Field

[0001] This invention relates to the field of elderly care robot technology, and in particular to a medical monitoring elderly care robot. Background Technology

[0002] How to provide better medical care and life security for the elderly has become an important issue of social concern. In particular, the demand for management of common chronic diseases, emotional support and daily life care for the elderly is increasing. Advances in modern technology have enabled robots to be gradually applied to the field of elderly care. Through sensors, data analysis and automatic control technology, real-time monitoring of the health status of the elderly, environmental regulation and personalized care support can be achieved, which provides a broad application prospect for the intelligent development of elderly care services.

[0003] Currently available elderly care robots typically possess only limited functions, such as simple health monitoring or assisted care. They are unable to comprehensively manage the complex health needs of the elderly, especially in health monitoring, where existing technologies struggle to simultaneously integrate physiological, environmental, and emotional data for accurate health assessments. Furthermore, they lack effective monitoring of indoor air quality, making it impossible to adjust the air environment in a timely manner based on the elderly's health status. In addition, emotional care is limited in the application of elderly care robots; emotional assessment accuracy is insufficient, intervention measures are not flexible enough, and they cannot provide personalized intervention plans in a timely manner based on the elderly's emotional fluctuations. These shortcomings limit the effectiveness of elderly care robots in improving the quality of life and care outcomes for the elderly.

[0004] This invention aims to address the shortcomings of existing technologies by proposing a medical monitoring-type elderly care robot that can automatically adjust the intensity and frequency of nursing actions to ensure that the elderly receive precise and comfortable care in their daily lives, thereby improving their quality of life and reducing their care burden. Summary of the Invention

[0005] This invention provides a medical monitoring-type elderly care robot.

[0006] A medical monitoring-type elderly care robot includes a physiological monitoring unit, an air particulate matter regulation and health feedback unit, an emotion and health integration unit, and a nursing assistance unit, wherein;

[0007] The physiological monitoring unit collects physiological data of the elderly in real time, including heart rate, blood pressure, blood oxygen, body temperature and respiratory rate;

[0008] The air particulate control unit detects pollutants and biological particles in the air, analyzes air quality in real time, and automatically adjusts the air purification equipment inside the robot or the external ventilation device based on the concentration of air particles and the physiological data of the elderly. Specifically, it includes:

[0009] Air quality monitoring: Real-time detection of pollutants and biological particles in the air, including bacteria, viruses, allergens, PM2.5 and volatile organic compounds (VOCs), generating air particulate concentration data;

[0010] Health impact assessment: Based on the generated air particulate concentration data and combined with the physiological data of the elderly, a correlation analysis is conducted through a health risk assessment model to assess the potential impact of air pollutants on the health of the elderly.

[0011] Air purification and environmental control: Based on the results of health impact assessment, automatically adjust the air purification equipment inside the robot or wirelessly control the external ventilation equipment;

[0012] The emotional and health fusion unit analyzes the mental health of the elderly by integrating facial expression recognition, voice analysis, and physiological data, and automatically generates emotional intervention plans based on the analysis results, including playing soothing music, adjusting lighting, or reminding caregivers to pay attention.

[0013] The nursing assistance unit includes a robotic arm and a voice control system to assist the elderly in daily activities, including eating, dressing, and exercise rehabilitation training. It also automatically adjusts the strength and frequency of the assistive movements based on feedback data from the physiological monitoring unit and the emotion and health integration unit.

[0014] Optionally, the physiological monitoring unit includes:

[0015] Heart rate monitoring: A heart rate sensor using photoplethysmography (PPG) is worn on the wrist or fingertip of the elderly to collect heart rate data in real time.

[0016] Blood pressure monitoring: Using a non-invasive inflatable cuff and pressure sensor, it is strapped to the upper arm or wrist of the elderly and blood pressure changes are monitored in real time through oscillation.

[0017] Blood oxygen monitoring: An optical blood oxygen sensor, mounted on a finger clip or earlobe clip, uses infrared and visible light transmission technology to measure the blood oxygen saturation of the elderly in real time.

[0018] Body temperature monitoring: Non-contact infrared body temperature sensors or patch temperature sensors are used and placed on the forehead, ear canal or skin surface of the elderly to obtain body temperature data in real time;

[0019] Respiratory rate monitoring: Real-time monitoring of respiratory rate and breathing pattern in the elderly using chest strap respiratory sensors, piezoelectric sensors, or radar-based non-contact sensors.

[0020] Optionally, the air quality detection includes:

[0021] Particulate matter detection: Detects particulate matter in the air, including PM2.5 and PM10, using laser scattering or optical detection technology to monitor particulate matter concentration in real time;

[0022] Volatile organic compound detection: Volatile organic compounds (VOCs) in the air are detected using electrochemical or semiconductor gas sensors, and gas concentration data is generated in real time;

[0023] Biological particle detection: Based on optical particle counting, it detects biological particles in the air, including bacteria, viruses and allergens, and analyzes their quantity and type in real time;

[0024] Multimodal data fusion: This involves comprehensively analyzing collected pollutant and biological particulate data to generate airborne particulate concentration data, represented as follows:

[0025] C air =w PM ·P PM +w VOC ·V VOC +w bio ·B bio ;

[0026] Among them, C air For the fused air particulate concentration data, P PM For PM2.5 or PM10 particulate matter data, V VOC For volatile organic compound data, B bio For biological microparticle data (such as bacteria, viruses, etc.), w PM 、w VOC 、w bio These are the corresponding weights.

[0027] Optionally, the health risk assessment model employs a Gaussian mixture model (GMM), which includes:

[0028] Data standardization: Standardize air particulate concentration data and physiological data;

[0029] The basic structure of a Gaussian mixture model: A Gaussian mixture model is used to model standardized air particle concentration data and physiological data, represented as follows:

[0030]

[0031] Where X represents multidimensional input data, including standardized airborne particulate concentration data and physiological data. It is the k-th Gaussian distribution with a mean of μ. k The covariance matrix is ​​∑ k , π kθ is the weight of the k-th mixture component, θ contains all model parameters, and K is the total number of Gaussian components (different potential health states, including healthy, mild risk, moderate risk, and severe risk).

[0032] Introducing personalized weights for health status: Introducing personalized weights for health status, dynamically adjusting the weights of different health risk statuses based on the different sensitivities of the elderly to air pollution (e.g., patients with chronic diseases are more susceptible to the effects of air pollution);

[0033] Dynamic parameter updates: To capture the impact of air pollutants on the health of the elderly, a time dimension is introduced to dynamically update the model parameters. The impact of historical air quality on health is smoothed by using time-weighted averaging.

[0034] Health risk probability output: Based on the dynamically updated Gaussian mixture model, the probability of each health state is output, representing the potential health risk of air pollution to the elderly.

[0035] Optionally, the air purification and environmental control include:

[0036] Air purification control: By receiving the results of a health impact assessment, the air purification equipment inside the robot is automatically adjusted, including adjusting the fan speed, filter working mode, and purification intensity;

[0037] External ventilation control: Connects to external intelligent ventilation equipment via wireless communication, and remotely controls the opening, closing, and ventilation intensity adjustment of the external intelligent ventilation equipment based on the results of health impact assessment, thereby achieving dynamic optimization of indoor air circulation and quality.

[0038] Optionally, the emotion and health integration unit includes:

[0039] Facial expression recognition: The system captures the facial expressions of the elderly through a camera and uses a convolutional neural network (CNN) to analyze facial features in real time to identify the elderly’s emotional state, including happiness, sadness, anger, and anxiety.

[0040] Voice emotion analysis: Collects the voice signals of the elderly through a microphone, assesses the emotional information in the voice, including tone, speech rate, and volume, and judges the emotional state of the elderly.

[0041] Data fusion and emotional intervention: The emotional state of facial expression recognition and voice emotion analysis is fused with physiological data to comprehensively analyze the mental health status of the elderly. Based on the analysis results, an emotional intervention plan is automatically generated, including playing soothing music, adjusting ambient lighting, or reminding caregivers to pay attention.

[0042] Optionally, the facial expression recognition includes:

[0043] Camera capture: Captures facial images of the elderly in real time using a camera;

[0044] Facial feature extraction: Real-time analysis of facial features based on convolutional neural networks (CNN) to extract feature vectors of key facial points, including the eyes, corners of the mouth, and eyebrow areas;

[0045] Emotion classification: The extracted facial feature vectors are input into a convolutional neural network for classification to identify the emotional states of the elderly, including happiness, sadness, anger, and anxiety.

[0046] Optionally, the voice emotion analysis includes:

[0047] Voice acquisition: Real-time acquisition of elderly people's voice signals via microphone;

[0048] Speech feature extraction: Speech signal processing algorithms are used to extract emotional features from speech, including intonation (fundamental frequency F0), speech rate and volume, which are used to determine the emotional information in speech.

[0049] Emotion classification: The extracted emotional features are input into the support vector machine (SVM) model, which outputs the emotional state of the elderly, including the emotional categories of happiness, sadness, anger, and anxiety.

[0050] Optionally, the data fusion and emotional intervention include:

[0051] Multimodal data fusion: The emotional state of facial expression recognition, the emotional state of voice emotion analysis, and physiological data are fused and processed, and a weighted average method is used to generate the current mental health index H of the elderly.

[0052] Mental health status assessment: Based on the integrated mental health index H, the mental health status of the elderly is assessed using a predefined threshold range and divided into different emotional risk levels, including low risk, medium risk and high risk.

[0053] Emotional intervention: Based on the assessment results, an emotional intervention plan is automatically generated, including playing soothing music, adjusting ambient lighting, or reminding nursing staff to pay attention.

[0054] Optionally, the nursing assistance unit includes:

[0055] Robotic arm assistance: Through data interaction with the physiological monitoring unit and the emotion and health integration unit, it receives real-time physiological data and psychological health feedback information of the elderly, and assists the elderly in daily activities, including feeding assistance, dressing assistance and exercise rehabilitation training.

[0056] Voice control system: Controls the robotic arm to perform auxiliary tasks through voice commands, and adjusts the intensity and frequency of the robotic arm's movements in real time by combining feedback data from the physiological monitoring unit and the emotion and health fusion unit.

[0057] The beneficial effects of this invention are:

[0058] This invention uses a physiological monitoring unit to collect real-time data on the elderly person's heart rate, blood pressure, blood oxygen, body temperature, and respiratory rate, ensuring a comprehensive understanding of their physical condition. At the same time, the air particulate control unit can accurately monitor pollutants and biological particles in the air and automatically adjust air purification equipment or external ventilation devices according to a health impact assessment model to ensure that indoor air quality meets health standards, prevent respiratory diseases, and provide a safe and healthy living environment for the elderly.

[0059] This invention integrates facial expressions, voice emotion analysis, and physiological data through an emotion and health fusion unit to comprehensively assess the mental health status of the elderly. Based on threshold ranges set from historical data, the system can dynamically determine the emotional risk level of the elderly and automatically generate emotional intervention plans, such as playing soothing music, adjusting lighting, or reminding caregivers to intervene. This provides timely psychological support and emotional care for the elderly, improves their quality of life, and effectively prevents mental health problems.

[0060] This invention assists the elderly in completing daily activities, including eating, dressing, and exercise rehabilitation training, through a robotic arm and voice control system. By combining physiological monitoring and emotional feedback, the robotic arm's movement strength and frequency are intelligently adjusted to achieve personalized care support. This intelligent and automated care solution not only improves the comfort and safety of the elderly but also reduces the workload of caregivers, providing the elderly with more convenient and efficient life care management. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of the robot functional units according to an embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of the air particulate control unit according to an embodiment of the present invention. Detailed Implementation

[0064] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0065] It should be noted that the use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0066] Generally, terms can be understood at least partly from their use in context. For example, depending at least partly on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in a singular sense, or a combination of features, structures, or characteristics in a plural sense. Additionally, the term "based on" can be understood not necessarily to convey an exclusive set of factors, but rather, alternatively, depending at least partly on the context, to allow for the presence of other factors that are not necessarily explicitly described.

[0067] like Figures 1-2 As shown, a medical monitoring-type elderly care robot includes a physiological monitoring unit, an air particulate matter regulation and health feedback unit, an emotion and health integration unit, and a nursing assistance unit, wherein;

[0068] The physiological monitoring unit collects physiological data of the elderly in real time, including heart rate, blood pressure, blood oxygen, body temperature and respiratory rate;

[0069] The air particulate control unit detects pollutants and biological particles in the air, analyzes air quality in real time, and automatically adjusts the air purification equipment inside the robot or the external ventilation device based on the concentration of airborne particles and the physiological data of the elderly, ensuring healthy indoor air and preventing damage to or infection of the elderly's respiratory system. Specifically, this includes:

[0070] Air quality monitoring: Real-time detection of pollutants and biological particles in the air, including bacteria, viruses, allergens, PM2.5 and volatile organic compounds (VOCs), generating air particulate concentration data;

[0071] Health impact assessment: Based on the generated air particulate concentration data and combined with the physiological data of the elderly, a correlation analysis is conducted through a health risk assessment model to assess the potential impact of air pollutants on the health of the elderly.

[0072] Air purification and environmental control: Based on the results of health impact assessment, the air purification equipment inside the robot body is automatically adjusted or the external ventilation equipment is wirelessly controlled to ensure that the air quality meets safety standards and optimize the living environment for the elderly.

[0073] The Emotion and Health Integration Unit analyzes the mental health of the elderly by integrating facial expression recognition, voice analysis, and physiological data, and automatically generates emotional intervention plans based on the analysis results, including playing soothing music, adjusting lighting, or reminding caregivers to pay attention.

[0074] The nursing assistance unit includes a robotic arm and a voice control system to assist the elderly in daily activities, including eating, dressing, and exercise rehabilitation training. Based on feedback data from the physiological monitoring unit and the emotion and health integration unit, it automatically adjusts the strength and frequency of the assistive movements to ensure personalized assistance support.

[0075] Through the above, the system achieves automated management of comprehensive health monitoring, environmental control, and personalized care for the elderly. It can not only dynamically adjust air quality and daily care plans based on the elderly’s real-time health status, but also provide psychological support through intelligent emotion analysis, ensuring that the elderly live in a safe, healthy, and comfortable environment.

[0076] The physiological monitoring unit includes:

[0077] Heart rate monitoring: A heart rate sensor using photoplethysmography (PPG) is worn on the wrist or fingertip of the elderly to collect heart rate data in real time.

[0078] Blood pressure monitoring: Using a non-invasive inflatable cuff and pressure sensor, it is strapped to the upper arm or wrist of the elderly and blood pressure changes are monitored in real time through oscillation.

[0079] Blood oxygen monitoring: An optical blood oxygen sensor, mounted on a finger clip or earlobe clip, uses infrared and visible light transmission technology to measure the blood oxygen saturation of the elderly in real time.

[0080] Body temperature monitoring: Non-contact infrared body temperature sensors or patch temperature sensors are used and placed on the forehead, ear canal or skin surface of the elderly to obtain body temperature data in real time;

[0081] Respiratory rate monitoring: Real-time monitoring of respiratory rate and breathing pattern in the elderly using chest strap respiratory sensors, piezoelectric sensors or radar non-contact sensors;

[0082] By implementing the above measures, we can ensure continuous tracking and early warning of abnormalities in the health status of the elderly, improve the accuracy and comfort of monitoring, effectively avoid the inconvenience of traditional monitoring methods, provide more personalized health management and timely intervention, and safeguard the safety of the elderly.

[0083] Air quality testing includes:

[0084] Particulate matter detection: Detects particulate matter in the air, including PM2.5 and PM10, using laser scattering or optical detection technology to monitor particulate matter concentration in real time;

[0085] Volatile organic compound detection: Volatile organic compounds (VOCs) in the air are detected using electrochemical or semiconductor gas sensors, and gas concentration data is generated in real time;

[0086] Biological particle detection: Based on optical particle counting, it detects biological particles in the air, including bacteria, viruses and allergens, and analyzes their quantity and type in real time;

[0087] Multimodal data fusion: This involves comprehensively analyzing collected pollutant and biological particulate data to generate airborne particulate concentration data, represented as follows:

[0088] C air =w PM ·P PM +w VOC ·V VOC +w bio ·B bio ;

[0089] Among them, C air For the fused air particulate concentration data, P PM For PM2.5 or PM10 particulate matter data, V VOC For volatile organic compound data, B bio For biological microparticle data (such as bacteria, viruses, etc.), w PM 、w VOC 、w bio These are the corresponding weights, w PM +w VOC +w bio =1;

[0090] weight w PM 、w VOC 、w bio The settings and possible values ​​are as follows:

[0091] weight w PM The value is relatively high because particulate matter such as PM2.5 has a significant impact on the respiratory system; therefore, the value is set at 0.5.

[0092] weight w VOC The value is low because it detects volatile organic compounds, but long-term exposure may affect the immune and nervous systems; therefore, it is set at 0.2.

[0093] weight wbio The value is relatively high, especially when detecting biological particles such as viruses and bacteria that have an important impact on respiratory health, with a value of 0.3.

[0094] The above information allows for a comprehensive assessment of the potential threats that air quality poses to the respiratory, immune, and long-term health of the elderly. This enables precise environmental monitoring and personalized health protection, effectively improving the reliability and adaptability of air quality testing and ensuring that the elderly can live in a safe air environment.

[0095] The health risk assessment model uses a Gaussian Mixture Model (GMM), which includes:

[0096] Data standardization: Air particulate concentration data and physiological data are standardized to ensure that input data from different dimensions have the same scale, represented as follows:

[0097]

[0098] Where X is the raw value of air particulate concentration data or physiological data, μ X It is the mean of the data, σ X X is the standard deviation of the data, and X′ is the standardized data.

[0099] The basic structure of a Gaussian mixture model: A Gaussian mixture model is used to model standardized air particle concentration data and physiological data, represented as follows:

[0100]

[0101] Where X represents multidimensional input data, including standardized airborne particulate concentration data and physiological data. It is the k-th Gaussian distribution with a mean of μ. k The covariance matrix is ​​∑ k , π k It is the weight of the k-th mixture component, and satisfies θ contains all model parameters, and K is the total number of Gaussian components (different potential health states, including healthy, low risk, moderate risk, and high risk).

[0102] Introducing personalized weights for health status: Personalized weights for health status are introduced, dynamically adjusting the weights of different health risk states based on the varying sensitivities of older adults to air pollution (e.g., those with chronic diseases are more susceptible to the effects of air pollution), thereby improving the accuracy of the assessment. This is represented as follows:

[0103] π k '=π k ·f(H);

[0104] Where, πk ′ is the adjusted weight of the k-th mixture component, and f(H) is a function used to adjust the weights based on the physiological condition H of the elderly (such as history of chronic diseases, respiratory diseases, etc.).

[0105] The function f(H) is calculated using a weighted average of different health factors, and is expressed as:

[0106] f(H) = 1 + w1·I respiratory +w2·I chronic +w3·I immune ;

[0107] Where w1, w2, and w3 are the weight coefficients of different health factors, representing the degree of influence of these health factors on risk sensitivity, and can be determined through data training. respiratory It is a binary indicator variable representing whether an elderly person suffers from respiratory diseases (such as asthma, chronic obstructive pulmonary disease, etc.). If such a disease is present, I... respiratory =1, otherwise 0, I chronic It is a binary indicator variable representing whether an elderly person has a history of chronic diseases (such as diabetes, hypertension, etc.). When a chronic disease is present, I... chronic =1, otherwise 0, I immune It is a binary indicator that represents the state of an older adult's immune system (e.g., whether they have weakened immunity). If their immunity is weakened, I... immune =1, otherwise 0;

[0108] Dynamic parameter updates: To capture the impact of air pollutants on the health of the elderly, a time dimension is introduced, and the model parameters are dynamically updated. The impact of historical air quality on health is smoothed using a time-weighted average, as shown below:

[0109]

[0110] in, It is the mean after updating at time t+1. This is the updated covariance matrix, where α is the time weighting coefficient, controlling the balance between the influence of historical and current data, and X... t The input data (air particle concentration and physiological data) is at time t. It is the mean of the k-th Gaussian distribution at time t. It is the covariance matrix of the k-th Gaussian distribution at time t. It is the product of the deviations between the input data and the mean, used to update the covariance matrix;

[0111] Health risk probability output: Based on the dynamically updated Gaussian mixture model, the probability of each health state is output, representing the potential health risk of air pollution to the elderly, expressed as:

[0112]

[0113] Among them, P(H k |X) represents the k-th health state of the elderly person given air particulate concentration data and physiological data X. k The probability of (health, low risk, moderate risk, high risk), π k ′ represents the weight of the k-th Gaussian component after personalized adjustment. Let be the probability density function of the k-th Gaussian distribution. The denominator is the weighted sum of probabilities of all health states, used to standardize the probability distribution and ensure that the sum of probabilities of each health state is 1.

[0114] The above methods enable a more accurate assessment of the potential health risks posed by air pollution to the elderly. By introducing personalized weights based on health status, the model dynamically adjusts these weights according to the different physiological conditions of the elderly (such as respiratory diseases or weakened immunity), thereby providing a personalized health risk assessment. At the same time, through dynamic parameter updates over time, the model can capture the long-term cumulative effects of air pollution, improving the sensitivity and prediction accuracy of health risks, enhancing the personalization of health management, and providing the elderly with more timely and effective health warnings.

[0115] Air purification and environmental control include:

[0116] Air purification control: By receiving the results of a health impact assessment, the robot automatically adjusts the air purification equipment inside, including adjusting the fan speed, filter working mode, and purification intensity, to ensure that the air quality meets the set health standards.

[0117] External ventilation control: Connects to external intelligent ventilation equipment via wireless communication, and remotely controls the opening, closing, and ventilation intensity adjustment of the external intelligent ventilation equipment based on the results of health impact assessment, thereby achieving dynamic optimization of indoor air circulation and quality;

[0118] The above-mentioned methods can provide personalized air management solutions to ensure that the elderly are always in a safe and healthy air environment regardless of their health conditions. This automated and personalized control method not only improves the accuracy of environmental management but also reduces the need for manual intervention, greatly enhancing the intelligence and convenience of elderly care.

[0119] The Emotional and Health Integration Unit includes:

[0120] Facial expression recognition: The system captures the facial expressions of the elderly through a camera and uses a convolutional neural network (CNN) to analyze facial features in real time to identify the elderly’s emotional state, including happiness, sadness, anger, and anxiety.

[0121] Voice emotion analysis: Collects the voice signals of the elderly through a microphone, assesses the emotional information in the voice, including tone, speech rate, and volume, and judges the emotional state of the elderly.

[0122] Data fusion and emotional intervention: The emotional state of facial expression recognition and voice emotion analysis is fused with physiological data to comprehensively analyze the mental health status of the elderly. Based on the analysis results, an emotional intervention plan is automatically generated, including playing soothing music, adjusting ambient lighting, or reminding caregivers to pay attention.

[0123] Through the above methods, we can perceive the emotional fluctuations of the elderly in real time, improve the accuracy of assessments, and automatically generate personalized emotional intervention plans that can effectively improve the emotions of the elderly and enhance their quality of life. Through intelligent and personalized emotional monitoring and intervention, we can not only improve the efficiency of elderly care, but also prevent mental health problems in a timely manner and provide more comprehensive care.

[0124] Facial expression recognition includes:

[0125] Camera capture: Captures facial images of the elderly in real time using a camera;

[0126] Facial Feature Extraction: Real-time analysis of facial features is performed based on a Convolutional Neural Network (CNN) to extract feature vectors from key facial points, including the eyes, corners of the mouth, and eyebrow regions, represented as follows:

[0127]

[0128] Among them, f l (x, y) represents the output feature map value of the l-th convolutional layer at position (x, y). l-1 (x+i,y+j are the input image pixel values ​​at position (x+i,y+j) of layer l-1, K l (i,j) is the weight matrix of the convolution kernel, with a size of m×n, b l is the bias term of the convolutional layer, m is the size of the convolutional kernel in the vertical direction, and n is the size of the convolutional kernel in the horizontal direction;

[0129] a l (x,y)=max(0,f l (x,y));

[0130] Among them, a l (x,y) is the activation value of the l-th convolutional layer at position (x,y);

[0131]

[0132] Among them, P l(x, y) represents the feature values ​​output by the pooling layer, a l (x·p+i, y·p+j) represents the activation values ​​within the pooling window, which are within the range of p×p and are the output values ​​of the convolutional layer provided by ReLU. p is the size of the pooling window, which is 2×2.

[0133] Emotion Classification: The extracted facial feature vectors are input into a convolutional neural network for classification to identify the emotional states of the elderly, including happiness, sadness, anger, and anxiety, represented as follows:

[0134]

[0135] Among them, z k W is the activation value for the k-th emotional state. ki For connection weights, v represents the weights between the i-th feature and the k-th class. i The flattened feature vector is the output from the previous convolution and pooling processes, b k For bias terms;

[0136]

[0137] Where P(y=k|x) is the probability that input x is classified as the k-th sentiment class, and z k is the activation value of the fully connected layer, used for the k-th emotional state, where K is the total number of emotional states;

[0138] y * =argmax k {P(y=k|x)};

[0139] Among them, y * To determine the final emotional state category, the emotional category with the highest probability is selected as the recognition result.

[0140] Based on the above, convolutional layers can be used to effectively extract key facial features, such as eyes, corners of the mouth, and eyebrows. Fully connected layers and the Softmax function can be used to accurately identify the emotional states of the elderly, including happiness, sadness, anger, and anxiety. It has efficient, real-time, and non-invasive emotion recognition capabilities, and can provide high-precision emotion monitoring without disturbing the elderly. This helps to adjust care plans in a timely manner and improve the quality of life of the elderly.

[0141] Voice sentiment analysis includes:

[0142] Voice capture: The voice signals of the elderly are captured in real time through a microphone to ensure the clarity and integrity of the voice data;

[0143] Speech feature extraction: Speech signal processing algorithms are used to extract emotional features from speech, including intonation (fundamental frequency F0), speech rate, and volume, which are used to determine the emotional information in the speech, as follows:

[0144]

[0145] Where F0(t) is the fundamental frequency (pitch) at time t, and T0(t) is the period at time t;

[0146]

[0147] Where S is the speech rate, representing the number of words spoken per unit time, and N... words It is the number of words in the audio segment, T speech It refers to the duration of the voice message;

[0148]

[0149] Where E is the energy (volume) of the speech signal, x[n] is the sampled speech signal value, and N is the number of sampling points of the speech signal;

[0150] Emotion Classification: The extracted emotional features are input into a Support Vector Machine (SVM) model, which outputs the emotional state of the elderly, including emotion categories such as happiness, sadness, anger, and anxiety, represented as follows:

[0151]

[0152] Where f(x) is the classification function, x i Here, x is the training sample, x is the input speech feature vector, and α is the input speech feature vector. i and y i It is a Lagrange multiplier and label, K(x) i (x) is the kernel function used to calculate the similarity between features, b is the bias term of the classifier, and M is the number of training samples;

[0153] The above methods enable real-time monitoring of emotional changes without visual contact with the elderly. By analyzing multi-dimensional voice features, more nuanced and accurate emotional assessments can be provided, offering an intelligent and real-time solution for emotional care and mental health monitoring of the elderly. This helps to detect emotional fluctuations in a timely manner and take appropriate interventions.

[0154] Data fusion and emotional intervention include:

[0155] Multimodal data fusion: This method integrates emotional states from facial expression recognition, emotional states from voice emotion analysis, and physiological data. A weighted average is then used to generate the current mental health index H for the elderly, expressed as:

[0156] H = w4·E facial +w5·E vocal +w6·P physio ;

[0157] Where H is the mental health index, used to characterize the mental health status of the elderly, and E... facial Emotional states for facial expression recognition vocal For the emotional state of speech sentiment analysis, P physio The results represent a comprehensive evaluation of physiological data, with w4, w5, and w6 serving as weighting coefficients.

[0158] The comprehensive evaluation result of physiological data P physio Expressed as:

[0159]

[0160] Among them, P physio S is a comprehensive evaluation result of physiological data. i For the standardized score of the i-th physiological indicator (such as the score of heart rate, blood pressure, blood oxygen, etc.), w i,physio Let be the weight of the i-th physiological indicator, representing the impact of the physiological indicator on health assessment, and p be the total number of physiological indicators.

[0161] The weight w of the i-th physiological indicator i,physio The weighting is based on the degree of influence of each physiological indicator on the mental health and mood fluctuations of the elderly. Generally, indicators that are closely related to mood and stress, such as heart rate, blood pressure, and respiratory rate, are given higher weights, while the weights of body temperature and blood oxygen indicators may be lower. The weight of heart rate is 0.4, the weight of blood pressure is 0.3, the weight of respiratory rate is 0.1, the weight of body temperature is 0.05, and the weight of blood oxygen indicator is 0.15.

[0162] The specific basis and values ​​for the weighting coefficients w4, w5, and w6 include:

[0163] The weighting coefficient w4 is set based on the importance of facial expressions in reflecting the emotional state of the elderly. If the elderly have obvious facial expressions, facial expression recognition may have a higher weight, and the value is 0.4.

[0164] The weighting factor w5 is set based on the fact that it can capture changes in the tone, speed and volume of the elderly, reflecting more subtle emotions. For some elderly people, their voice may reflect emotional fluctuations better than facial expressions. The value is 0.3.

[0165] The weighting coefficient w6 is set based on the fact that it has a direct impact on mental health, especially since heart rate and blood pressure can reflect emotional fluctuations. The weight setting depends on the importance of physiological health status. When elderly people have chronic diseases or cardiovascular problems, the weight of physiological data will increase, and the value is 0.3.

[0166] Mental health status assessment: Based on the integrated mental health index H, the mental health status of the elderly is assessed using a predefined threshold range and divided into different emotional risk levels, including low risk, medium risk and high risk.

[0167] Emotional intervention: Based on the assessment results, an emotional intervention plan is automatically generated, including playing soothing music, adjusting ambient lighting, or reminding nursing staff to pay attention;

[0168] The predefined threshold ranges are set based on historical data and include:

[0169] Data collection: Collect historical health data of the elderly, including facial expression recognition, voice emotion analysis and physiological data, and generate a corresponding mental health index H for each data point;

[0170] Data statistical analysis: The collected health indices are statistically analyzed to calculate the distribution of the mental health index H, and risk levels are classified by percentiles;

[0171] Percentile threshold setting: Based on historical data percentiles, health indices below the 25th percentile are considered low risk, those between the 25th and 75th percentiles are considered medium risk, and those above the 75th percentile are considered high risk, expressed as:

[0172]

[0173] Where H is the mental health index, generated by the fusion of multimodal data; T1 is the 25th percentile in historical data, corresponding to the upper limit of low risk; and T2 is the 75th percentile in historical data, corresponding to the lower limit of high risk.

[0174] The above methods enable accurate assessment of the mental health status of the elderly and automatic generation of personalized intervention plans. The use of multimodal data fusion improves the accuracy of the assessment and avoids the bias that may be caused by a single data source. At the same time, the risk threshold setting based on historical data can dynamically adapt to the emotional fluctuations of individual elderly people. It can not only achieve real-time emotional monitoring, but also provide personalized emotional intervention plans according to different risk levels, thereby improving the mental health and quality of life of the elderly.

[0175] The nursing support unit includes:

[0176] Robotic arm assistance: Through data interaction with the physiological monitoring unit and the emotion and health integration unit, it receives real-time physiological data and psychological health feedback information of the elderly, and assists the elderly in daily activities, including feeding assistance, dressing assistance and exercise rehabilitation training.

[0177] Feeding assistance uses a robotic arm to deliver food based on the elderly person's hand movements and eating needs, adjusting the delivery speed and force to avoid operations that are too fast or too slow;

[0178] The dressing aid adjusts the force of the robotic arm to assist dressing based on the elderly person's hand dexterity and body posture, ensuring a smooth and comfortable dressing process.

[0179] Exercise rehabilitation training monitors the limb condition of the elderly, and the robotic arm automatically controls the operating force and frequency of the exercise rehabilitation equipment to ensure the safety and effectiveness of exercise training.

[0180] Voice control system: The elderly control the robotic arm to perform assistive tasks through voice commands. Combined with feedback data from the physiological monitoring unit and the emotion and health fusion unit, the intensity and frequency of the robotic arm's movements are adjusted in real time, as shown below:

[0181] F adjust =w7·P physio +w8·E emotion ;

[0182] Among them, F adjust P is used to adjust the force and frequency of the robotic arm's movements. physio E is the feedback value from the physiological monitoring unit (such as heart rate, blood pressure, etc.). emotion The feedback value for the emotional and health integration unit, with w7 and w8 as weighting coefficients;

[0183] The basis for setting the weighting coefficients w7 and w8, and their value range, specifically include:

[0184] The weighting coefficient w7 is set based on the fact that physiological feedback (such as heart rate, blood pressure, etc.) directly reflects the physical condition of the elderly. Especially when performing exercise rehabilitation or other physical activities, changes in heart rate and blood pressure are very important for adjusting assisted movements. If physiological monitoring data indicates that the elderly are in a state of high cardiovascular load, the system will prioritize adjusting the force and frequency of the robotic arm to reduce the physical stress on the elderly. Therefore, the weight of physiological feedback is relatively large, and the value is 0.7.

[0185] The weighting coefficient w8 is set based on the fact that emotional feedback reflects the psychological state of the elderly. If the elderly person shows signs of stress, anxiety, or discomfort, the system needs to adjust the assist force and movement frequency of the robotic arm according to the emotional state. The weight of emotional feedback will depend on the degree of influence of the elderly person's emotional fluctuations on task execution, and is set to 0.3.

[0186] Through the above, the strength and frequency of assistive movements can be intelligently adjusted to provide personalized nursing support for the elderly. It can perceive the elderly’s physical condition and emotional changes in real time and automatically adjust the execution of assistive tasks, such as eating, dressing and exercise rehabilitation training, to ensure that each operation is both safe and comfortable. The nursing assistance unit realizes dynamic adjustment according to individual needs, improves the comfort and quality of life of the elderly in their daily lives, and reduces the workload of caregivers.

[0187] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0188] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A medical monitoring-type elderly care robot, characterized in that, It includes a physiological monitoring unit, an air particulate control unit, an emotional and health integration unit, and a nursing support unit, among which; The physiological monitoring unit collects physiological data of the elderly in real time, including heart rate, blood pressure, blood oxygen, body temperature and respiratory rate; The air particulate control unit detects pollutants in the air, analyzes air quality in real time, and automatically adjusts the air purification equipment inside the robot or the external ventilation equipment based on the concentration of air particles and the physiological data of the elderly. Specifically, it includes: Air quality monitoring: Real-time detection of pollutants in the air, including bacteria, viruses, and allergens, generating data on the concentration of airborne particulate matter; Health impact assessment: Based on the generated air particulate concentration data and combined with the physiological data of the elderly, a correlation analysis is conducted through a health risk assessment model to assess the potential impact of air pollutants on the health of the elderly. Air purification and environmental control: Based on the results of health impact assessment, automatically adjust the air purification equipment inside the robot or wirelessly control the external ventilation equipment; The emotional and health fusion unit analyzes the mental health of the elderly by integrating facial expression recognition, voice analysis, and physiological data, and automatically generates emotional intervention plans based on the analysis results, including playing soothing music, adjusting lighting, or reminding caregivers to pay attention. The nursing assistance unit includes a robotic arm and a voice control system to assist the elderly in daily activities, including eating, dressing, and exercise rehabilitation training. It also automatically adjusts the strength and frequency of the assistive movements based on feedback data from the physiological monitoring unit and the emotion and health integration unit. The health risk assessment model adopts a Gaussian mixture model, which includes: Data standardization: Standardizing airborne particulate concentration data and physiological data. ； in, These are the raw values ​​of airborne particulate matter concentration data or physiological data. It is the mean of the data. It is the standard deviation of the data. It is standardized data; The basic structure of a Gaussian mixture model: A Gaussian mixture model is used to model standardized air particle concentration data and physiological data, represented as follows: ； in, It is multidimensional input data, including standardized airborne particulate concentration data or physiological data. It is the first A Gaussian distribution with a mean of , The covariance matrix is , It is the first The weights of each component in the mixture Includes all model parameters. It is the total number of Gaussian components; Introducing personalized weights for health status: Introducing personalized weights for health status, dynamically adjusting the weights of different health risk statuses based on the different sensitivities of the elderly to air pollution; Dynamic parameter updates: To capture the impact of air pollutants on the health of the elderly, a time dimension is introduced to dynamically update the model parameters. The impact of historical air quality on health is smoothed by using time-weighted averaging. Health risk probability output: Based on the dynamically updated Gaussian mixture model, the probability of each health state is output, representing the potential health risks of air pollution to the elderly; The emotion and health integration unit includes: Facial expression recognition: The system captures the facial expressions of the elderly through a camera, and uses a convolutional neural network to analyze facial features in real time to identify the elderly’s emotional state, including happiness, sadness, anger, and anxiety. Voice emotion analysis: Collects the voice signals of the elderly through a microphone, assesses the emotional information in the voice, including tone, speech rate, and volume, and judges the emotional state of the elderly. Data fusion and emotional intervention: The emotional state of facial expression recognition and voice emotion analysis is fused with physiological data to comprehensively analyze the mental health status of the elderly. Based on the analysis results, an emotional intervention plan is automatically generated, including playing soothing music, adjusting ambient lighting, or reminding caregivers to pay attention.

2. The medical monitoring-type elderly care robot according to claim 1, characterized in that, The physiological monitoring unit includes: Heart rate monitoring: A heart rate sensor using photoplethysmography is worn on the wrist or fingertip of the elderly to collect heart rate data in real time; Blood pressure monitoring: Using a non-invasive inflatable cuff and pressure sensor, it is strapped to the upper arm or wrist of the elderly and blood pressure changes are monitored in real time through oscillation. Blood oxygen monitoring: An optical blood oxygen sensor, mounted on a finger clip or earlobe clip, uses infrared and visible light transmission technology to measure the blood oxygen saturation of the elderly in real time. Body temperature monitoring: Non-contact infrared body temperature sensors or patch temperature sensors are used and placed on the forehead and ear canals of elderly people to obtain body temperature data in real time; Respiratory rate monitoring: Real-time monitoring of respiratory rate and breathing pattern in the elderly using chest strap respiratory sensors, piezoelectric sensors, or radar-based non-contact sensors.

3. The medical monitoring-type elderly care robot according to claim 1, characterized in that, The air quality detection includes: Particulate matter detection: Detects particulate matter in the air, including PM2.5 and PM10, using laser scattering or optical detection technology to monitor particulate matter concentration in real time; Volatile organic compound detection: Volatile organic compounds in the air are detected using electrochemical or semiconductor gas sensors, and gas concentration data is generated in real time; Biological particle detection: Based on optical particle counting, it detects biological particles in the air, including bacteria, viruses and allergens, and analyzes their quantity and type in real time; Multimodal data fusion: This involves comprehensively analyzing collected pollutant and biological particulate data to generate airborne particulate concentration data, represented as follows: ； in, The data represents the fused air particulate concentration. For PM2.5 or PM10 particulate matter data, Data for volatile organic compounds. For biological microparticle data, , , These are the corresponding weights.

4. The medical monitoring-type elderly care robot according to claim 3, characterized in that, The air purification and environmental control include: Air purification control: By receiving the results of a health impact assessment, the air purification equipment inside the robot is automatically adjusted, including adjusting the fan speed, filter working mode, and purification intensity; External ventilation control: Connects to external ventilation equipment via wireless communication and remotely controls the opening, closing, and ventilation intensity adjustment of the external intelligent ventilation equipment based on the results of health impact assessment, thereby achieving dynamic optimization of indoor air circulation and quality.

5. A medical monitoring-type elderly care robot according to claim 1, characterized in that, The facial expression recognition includes: Camera capture: Captures facial images of the elderly in real time using a camera; Facial feature extraction: Real-time analysis of facial features based on convolutional neural networks, extracting feature vectors of key facial points, including the eyes, corners of the mouth, and eyebrow areas; Emotion classification: Extracted facial key points are input into a convolutional neural network for classification to identify the emotional states of the elderly, including happiness, sadness, anger, and anxiety.

6. A medical monitoring-type elderly care robot according to claim 5, characterized in that, The voice emotion analysis includes: Voice acquisition: Real-time acquisition of elderly people's voice signals via microphone; Speech feature extraction: Speech signal processing algorithms are used to extract emotional features from speech, including intonation, speech rate and volume, which are used to determine the emotional information in speech. Emotion classification: The extracted emotional features are input into the support vector machine model, which outputs the emotional state of the elderly, including the emotional categories of happiness, sadness, anger, and anxiety.

7. A medical monitoring-type elderly care robot according to claim 6, characterized in that, The data fusion and sentiment intervention include: Multimodal data fusion: This method integrates emotional states from facial expression recognition, emotional states from voice emotion analysis, and physiological data, and uses a weighted average method to generate a comprehensive mental health index for the elderly. ; Mental health status assessment: based on the integrated mental health index. The mental health status of the elderly is determined by using predefined threshold ranges and divided into different emotional risk levels, including low risk, medium risk and high risk. Emotional intervention: Based on the assessment results, an emotional intervention plan is automatically generated, including playing soothing music, adjusting ambient lighting, or reminding nursing staff to pay attention.

8. A medical monitoring-type elderly care robot according to claim 1, characterized in that, The nursing support unit includes: Robotic arm assistance: Through data interaction with the physiological monitoring unit and the emotion and health integration unit, it receives real-time physiological data and psychological health feedback information of the elderly, and assists the elderly in daily activities, including feeding assistance, dressing assistance and exercise rehabilitation training. Voice control system: Controls the robotic arm to perform auxiliary tasks through voice commands, and adjusts the intensity and frequency of the robotic arm's movements in real time by combining feedback data from the physiological monitoring unit and the emotion and health fusion unit.

Images