A voice emergency rescue method and system for the elderly based on the Internet of Things

Abstract

The present invention relates to an elderly voice emergency rescue method based on the Internet of Things, belonging to the field of data processing technology specifically suitable for administrative, commercial, financial, management, supervision or prediction purposes, and the method comprises: when receiving a fuzzy signal of a voice for help, evenly dividing the emergency rescue voice segment of a set elderly patient on the time axis to obtain the amplitude of each signal; using a deep neural network model to intelligently identify the emergency rescue type coding value corresponding to the emergency rescue voice segment based on the Internet of Things data and the amplitude of each signal. The present invention also relates to an elderly voice emergency rescue system based on the Internet of Things; through the present invention, in view of the technical problem in the prior art that it is difficult to identify the emergency rescue type of the fuzzy voice emitted by the weak elderly patient, an artificial intelligence model can be introduced to use the Internet of Things data to perform intelligent analysis on the fuzzy elderly patient voice signal, thereby solving the above technical problem.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology specifically suitable for administrative, commercial, financial, management, supervisory or forecasting purposes, and in particular to an Internet of Things-based voice emergency rescue method and system for the elderly. Background Art

[0002] Elderly patients usually refer to the elderly group who have reached a certain age and have experienced degenerative changes in their physical functions and physiological states, usually those aged 60 and above. With age, the physical functions of the elderly gradually decline, including the cardiovascular system, respiratory system, digestive system, nervous system and other aspects. They often suffer from a variety of chronic diseases, such as hypertension, diabetes, coronary heart disease, stroke, etc., which not only affect the quality of life of elderly patients, but also may increase the risk of complications and death. In addition, elderly patients are also prone to skeletal system diseases such as osteoporosis and arthritis, as well as mental and psychological problems such as cognitive impairment and depression. Therefore, elderly patients are the patient group with the highest probability of hospitalization, and they are also the most difficult patient group to care for at the nurse workstation of the inpatient department.

[0003] For example, the Chinese invention patent with publication number CN105489217A proposes a method and system for voice recognition care. This voice recognition care system includes a host, at least one slave and a care unit. The slave is installed in the room, detects multiple sound signals existing simultaneously in the environment, and uses whether the sound signal has a fundamental frequency-overband structure to determine whether the sound signal is a speech signal or a non-speech signal; if the sound signal is a voice signal, the text content of the voice signal is identified to determine whether it contains one of the preset multiple distress messages, and if the sound signal is a non-speech signal, it is determined whether it is one of the preset multiple abnormal sounds; if the text content of the voice signal is a distress message, or the non-speech signal is an abnormal sound, the host sends a warning message to the care unit to notify the family or caregiver to come and check.

[0004] For example, the Chinese invention patent with publication number CN108280911A proposes an intelligent medical monitoring system for the elderly based on the Internet of Things, which belongs to the field of the Internet of Things. The system includes an aggregator, an intelligent monitoring device, an intelligent medicine box, a wireless transmission device and an intelligent door lock; the intelligent medicine box records the medication situation and reminds the elderly, and transmits it to the doctor. The doctor adjusts the medicine according to the physical condition of the elderly monitored by the intelligent monitoring device, and feeds back to the intelligent medicine box; if the aggregator determines that the elderly's physiological condition is abnormal, the emergency rescue protocol is activated, a temporary door lock password is generated, and sent to the intelligent door lock, and a distress message is sent to the hospital through the wireless transmission device; the doctor opens the intelligent door lock according to the received information, makes a comprehensive judgment based on the recorded medication situation and provides rescue. The present invention can provide a complete medication record and physical condition of the elderly in an emergency, improve the treatment efficiency; when the doctor rescues the elderly, there is no need to waste time on finding a place to live and opening the door lock, which increases the chance of survival of the elderly.

[0005] However, the above technical scheme is only limited to analyzing and judging the emergency rescue information of the elderly patients based on their on-site voice signals or other physiological data obtained based on the Internet of Things. In fact, some elderly patients are very weak due to respiratory or heart diseases, and it is difficult for them to accurately send out emergency rescue voice fragments that are easy to be recognized by voice. As a result, the elderly patients send out vague emergency rescue voice fragments that are difficult to be accurately recognized. Naturally, it is impossible to provide corresponding emergency rescue for the elderly patients' physical condition, which can easily delay the best time for rescue. Summary of the invention

[0006] In order to solve the technical problems in the prior art, the present invention provides a voice emergency help method and system for the elderly based on the Internet of Things. When the character string obtained by conventional voice content analysis of the elderly patient's voice signal cannot accurately and reliably express the corresponding emergency help information, an artificial intelligence model with a customized structure design is introduced to use various physiological data obtained based on the Internet of Things to perform intelligent analysis on the fuzzy elderly patient's voice signal to obtain the current emergency help type, so as to achieve the effect of effectively judging the elderly patient's purpose of help through fuzzy language, thereby facilitating the corresponding emergency rescue of the elderly patient's physical condition and improving the work effect and efficiency of the nurse station in the inpatient department.

[0007] According to a first aspect of the present invention, a method for elderly people to call for emergency help via voice based on the Internet of Things is provided, the method comprising:

[0008] Storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request;

[0009] Perform voice content analysis on a set elderly patient's emergency rescue voice clip obtained through a voice collection device connected to the Internet of Things, and send a clear distress voice signal when the character string obtained through analysis matches the semantics of a certain emergency rescue type; and send an ambiguous distress voice signal when the character string obtained through analysis does not match the semantics of any emergency rescue type;

[0010] When receiving the fuzzy distress voice signal, the emergency distress voice segment is evenly divided on the time axis to obtain the amplitudes of the signals corresponding to the evenly spaced moments on the time axis;

[0011] Obtaining various items of associated information of a set elderly patient, wherein the various items of associated information of the set elderly patient are the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, a character string corresponding to a standard paragraph, and a standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph;

[0012] The current blood pressure, current heart rate and current blood oxygen saturation of a set elderly patient are obtained from multiple physiological parameter sensors connected to the Internet of Things, and a deep neural network model is used to intelligently identify the emergency help type coding value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis, and the corresponding emergency help type is mapped based on the intelligently predicted emergency help type coding value to output as the current emergency help type.

[0013] According to a second aspect of the present invention, there is provided an elderly voice emergency rescue system based on the Internet of Things, the system comprising a memory and one or more processors, the memory storing a computer program, the computer program being configured to be executed by the one or more processors to complete the following steps:

[0014] Storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request;

[0015] Perform voice content analysis on a set elderly patient's emergency rescue voice clip obtained through a voice collection device connected to the Internet of Things, and send a clear distress voice signal when the character string obtained through analysis matches the semantics of a certain emergency rescue type; and send an ambiguous distress voice signal when the character string obtained through analysis does not match the semantics of any emergency rescue type;

[0016] When receiving the fuzzy distress voice signal, the emergency distress voice segment is evenly divided on the time axis to obtain the amplitudes of the signals corresponding to the evenly spaced moments on the time axis;

[0017] Obtaining various items of associated information of a set elderly patient, wherein the various items of associated information of the set elderly patient are the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, a character string corresponding to a standard paragraph, and a standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph;

[0018] The current blood pressure, current heart rate and current blood oxygen saturation of a set elderly patient are obtained from multiple physiological parameter sensors connected to the Internet of Things, and a deep neural network model is used to intelligently identify the emergency help type coding value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis, and the corresponding emergency help type is mapped based on the intelligently predicted emergency help type coding value to output as the current emergency help type.

[0019] According to a third aspect of the present invention, there is provided an elderly voice emergency rescue system based on the Internet of Things, the system comprising:

[0020] A data storage device, used to store different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request;

[0021] A content judgment device is used to perform voice content analysis on a set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things, and when the character string obtained through analysis matches the semantics of a certain emergency rescue type, send a clear distress voice signal; and when the character string obtained through analysis does not match the semantics of any emergency rescue type, send an ambiguous distress voice signal;

[0022] A segmentation and collection device connected to the content judgment device is used to evenly segment the emergency rescue voice segment on the time axis when receiving the fuzzy distress voice signal, so as to obtain the amplitudes of the signals corresponding to the evenly spaced moments on the time axis;

[0023] An information parsing device is used to obtain various related information of a set elderly patient, wherein the various related information of the set elderly patient is the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, a character string corresponding to a standard paragraph, and a standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph;

[0024] A request identification device is respectively connected to the content judgment device, the segmentation and acquisition device, the information analysis device and the data storage device, and is used to obtain the current blood pressure, current heart rate and current blood oxygen saturation of a set elderly patient from multiple physiological parameter sensors connected to the Internet of Things, and use a deep neural network model to intelligently identify the emergency help type coding value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis, and map the corresponding emergency help type based on the intelligently predicted emergency help type coding value to output as the current emergency help type.

[0025] Compared with the prior art, the present invention has at least the following four main inventive features:

[0026] First invention point: When the character string obtained by conventional voice content analysis of the elderly patient's voice signal cannot accurately and reliably express the corresponding emergency help information, a customized artificial intelligence model is used to intelligently analyze the fuzzy elderly patient's voice signal based on multiple basic data selected in a targeted manner to obtain the current emergency help type, so as to achieve the effect of effectively judging the elderly patient's purpose of help through fuzzy language, and provide key data for the subsequent response of the nurse station and on-site rescue;

[0027] Second invention point: the deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer, and a single input layer, and the number of hidden layers is proportional to the set length of hospital stay for elderly patients, thereby completing model customization based on the number of emergency rescue types and the set length of hospital stay for elderly patients;

[0028] Third invention point: The multiple basic data selected in a targeted manner include the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient obtained from multiple physiological parameter sensors connected to the Internet of Things, the signal amplitudes corresponding to each time period evenly spaced on the time axis corresponding to the elderly patient's voice signal, and various related information of the set elderly patient. Specifically, the various related information of the set elderly patient is the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the elderly patient reading the standard paragraph;

[0029] The fourth invention point: in each learning execution of the deep neural network, the known emergency help type code value corresponding to a past emergency help voice clip issued by the set elderly patient is used as the single output data of the deep neural network, the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued a past emergency help voice clip, various related information of the set elderly patient, and the signal amplitudes corresponding to each time evenly spaced on the time axis corresponding to a past emergency help voice clip are used as input data of the deep neural network to execute this learning action, thereby ensuring the learning effect of each learning action of the deep neural network. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The embodiments of the present invention will be described below with reference to the accompanying drawings, wherein:

[0031] Figure 1 The present invention is a technical flow chart of the method and system for elderly people's voice emergency rescue based on the Internet of Things.

[0032] Figure 2 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 1 of the present invention.

[0033] Figure 3 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 2 of the present invention.

[0034] Figure 4 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 3 of the present invention.

[0035] Figure 5 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 4 of the present invention.

[0036] Figure 6The figure is a schematic diagram of the structure of an elderly voice emergency rescue system based on the Internet of Things according to Example 5 of the present invention.

[0037] Figure 7 The figure is a schematic diagram of the structure of an elderly voice emergency rescue system based on the Internet of Things according to Example 6 of the present invention. DETAILED DESCRIPTION

[0038] The Internet of Things is an "Internet of Things" covering everything in the world, based on the computer Internet, using technologies such as RFID and wireless data communication. In this network, objects (goods) can "communicate" with each other without human intervention. Its essence is to use radio frequency automatic identification (RFID) technology to achieve automatic identification of objects (goods) and interconnection and sharing of information through the computer Internet. The advent of the concept of the Internet of Things has broken the previous traditional thinking. In the past, the idea has always been to separate physical infrastructure from IT infrastructure, with airports, roads, and buildings on the one hand and data centers, personal computers, broadband, etc. on the other. In the era of the Internet of Things, reinforced concrete and cables will be integrated with chips and broadband into a unified infrastructure. In this sense, infrastructure is more like a new earth. Therefore, some industry insiders believe that the Internet of Things and smart grids are both organic components of the smart earth.

[0039] Regarding the difficulties of inpatient care, the care of elderly patients is more difficult. The multiple sensors arranged for elderly patients need to be aggregated, managed and utilized in a unified manner. At this time, relying on the multiple sensors arranged for the elderly patients, the Internet of Things is used to aggregate and uniformly manage and utilize the real-time data of multiple sensors, providing more real-time and accurate solutions for the care of elderly patients. This is an important breakthrough direction for smart medical care based on data processing for supervision or prediction purposes.

[0040] like Figure 1 As shown, a technical flow chart of the method and system for elderly people's voice emergency rescue based on the Internet of Things according to the present invention is given.

[0041] exist Figure 1 In the present invention, clear voice signals of elderly patients are directly analyzed through conventional voice content analysis to obtain corresponding emergency help information and provide it to the nearest nurse station, while fuzzy voice signals of elderly patients need to be analyzed through an intelligent analysis mechanism to obtain corresponding emergency help information and provide it to the nearest nurse station.

[0042] like Figure 1 As shown, the specific technical process of the present invention is as follows:

[0043] Technical process 1: Perform conventional voice content analysis on the elderly patient's voice signal at the bedside of the elderly patient. When the character string obtained through conventional voice content analysis cannot accurately and reliably express the corresponding emergency help information, the elderly patient's voice signal is judged to be a fuzzy voice signal to trigger the subsequent semantic recognition of the fuzzy voice signal based on the artificial intelligence model;

[0044] Technical process 2: Design a customized artificial intelligence model to perform semantic recognition of ambiguous elderly patient voice signals to determine the current emergency rescue type that the elderly patient actually wants to express;

[0045] For example, various emergency help types include bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request, and emergency transport request, and the current emergency help type that the elderly patient actually wants to express is one of them;

[0046] Specifically, the artificial intelligence model is a deep neural network model, and its structural customization is mainly reflected in the following aspects:

[0047] First, the deep neural network model is a deep neural network that has completed multiple learnings;

[0048] Secondly, the number of times the deep neural network is learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer. More importantly, the number of hidden layers is proportional to the set length of hospital stay for the elderly patient. In this way, model customization based on the number of emergency rescue types and the set length of hospital stay for the elderly patient is completed;

[0049] Finally, in each learning execution of the deep neural network, the known emergency rescue type code value corresponding to a certain emergency rescue voice segment issued by the set elderly patient in the past is used as the single output data of the deep neural network, and the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued an emergency rescue voice segment in the past, various related information of the set elderly patient, and the signal amplitudes corresponding to each time evenly spaced on the time axis corresponding to a certain emergency rescue voice segment issued in the past are used as input data of the deep neural network to execute this learning action, thereby ensuring the learning effect of each learning action of the deep neural network;

[0050] Technical process three: In order to perform semantic recognition of ambiguous elderly patients' speech signals, multiple basic data are fully and comprehensively selected to ensure the effectiveness and stability of semantic recognition results;

[0051] Specifically, the multiple basic data include the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient obtained from multiple physiological parameter sensors connected to the Internet of Things, the amplitudes of the signals corresponding to the evenly spaced moments on the time axis corresponding to the elderly patient's voice signal, and various related information of the set elderly patient;

[0052] More specifically, the various related information of the elderly patient is set to include the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the elderly patient reading the standard paragraph;

[0053] Technical process 4: Adopting a customized artificial intelligence model to fully and comprehensively select multiple basic data to intelligently analyze the emergency rescue type corresponding to the ambiguous voice signal of the elderly patient;

[0054] Technical process five: The emergency rescue type corresponding to the ambiguous voice signal of the elderly patient obtained by intelligent analysis is sent to the nearest nurse station in the inpatient department, and the real-time display is performed on the on-site control panel, thereby providing key data for the subsequent response of the nurse station and on-site rescue.

[0055] It can be seen that through the design of the above-mentioned multiple technical processes, even if the voice signal of the elderly patient is too vague and the corresponding emergency rescue type cannot be obtained through conventional voice content analysis, the artificial intelligence model with customized structure design can be used to accurately and reliably analyze the current emergency rescue type corresponding to the vague elderly patient's voice signal based on comprehensive and fully selected multiple basic data, thereby ensuring the speed and efficiency of nursing and rescue for elderly patients.

[0056] The key points of the present invention are: targeted identification of ambiguous elderly patient voice signals, structural customization of an artificial intelligence model for performing intelligent analysis of the current emergency rescue type corresponding to the ambiguous elderly patient voice signals, and targeted screening of multiple basic data for performing intelligent analysis.

[0057] Next, the method and system for elderly people's voice emergency rescue based on the Internet of Things of the present invention will be specifically described in the form of embodiments.

[0058] Example 1

[0059] Figure 2 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 1 of the present invention.

[0060] like Figure 2 As shown, the voice emergency rescue method for the elderly based on the Internet of Things includes the following specific steps:

[0061] Step S21: storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request;

[0062] For example, storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request respectively includes: the different emergency rescue type code values ​​are different binary code values;

[0063] Step S22: performing voice content analysis on the emergency help voice segment of the set elderly patient obtained through the voice collection device connected to the Internet of Things, and sending a clear help voice signal when the character string obtained through analysis matches the semantics of a certain emergency help type, and sending a vague help voice signal when the character string obtained through analysis does not match the semantics of any emergency help type;

[0064] Specifically, a voice content analysis is performed on a set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things, and when the character string obtained through analysis matches the semantics of a certain emergency rescue type, a clear distress voice signal is issued, and when the character string obtained through analysis does not match the semantics of any emergency rescue type, an ambiguous distress voice signal is issued, including: a mature voice recognition application on the market can be used to perform voice content analysis on the set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things;

[0065] Step S23: when receiving the fuzzy distress voice signal, evenly divide the emergency distress voice segment on the time axis to obtain the signal amplitudes corresponding to each time moment evenly spaced on the time axis;

[0066] For example, when receiving an ambiguous distress voice signal, evenly dividing the emergency distress voice segment on the time axis to obtain the signal amplitudes corresponding to each time evenly spaced on the time axis includes: obtaining a single signal amplitude of the emergency distress voice segment every 0.5 milliseconds on the time axis;

[0067] Step S24: obtaining various related information of the set elderly patient, wherein the various related information of the set elderly patient is the disease type, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph;

[0068] Specifically, obtaining various items of associated information of the set elderly patient, wherein the various items of associated information of the set elderly patient are the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph, including: a programmable logic device can be selected to obtain various items of associated information of the set elderly patient;

[0069] Step S25: obtaining the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from a plurality of physiological parameter sensors connected to the Internet of Things, using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various associated information of the set elderly patient and the signal amplitudes corresponding to each moment evenly spaced on the time axis, and mapping the corresponding emergency help type based on the intelligently predicted emergency help type code value to output as the current emergency help type;

[0070] For example, the MATLAB toolbox can be selected to complete the test and simulation of the data processing process of using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis;

[0071] The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the hospitalization duration of the set elderly patient;

[0072] For example, the number of times the deep neural network is learned is positively correlated with the number of types of different emergency rescue types, including: the greater the number of types of different emergency rescue types, the greater the number of times the deep neural network is learned;

[0073] And by way of example, the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the set length of hospitalization of the elderly patient, including: the set length of hospitalization of the elderly patient is 10 days, the number of hidden layers is 3, the set length of hospitalization of the elderly patient is 15 days, the number of hidden layers is 4, the set length of hospitalization of the elderly patient is 20 days, the number of hidden layers is 5, and so on;

[0074] The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the length of hospitalization of the set elderly patient. The deep neural network model includes: in each learning performed on the deep neural network, the emergency rescue type code value corresponding to a known emergency rescue voice segment issued by the set elderly patient in the past is used as the single output data of the deep neural network, the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued an emergency rescue voice segment in the past, various related information of the set elderly patient, and each signal amplitude corresponding to each time evenly spaced on the time axis corresponding to a past emergency rescue voice segment are used as input data of the deep neural network to perform this learning action;

[0075] Wherein, obtaining the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from multiple physiological parameter sensors connected to the Internet of Things includes: the multiple physiological parameter sensors connected to the Internet of Things are a blood pressure sensor, a heart rate sensor and a blood oxygen saturation sensor;

[0076] Among them, the voice content analysis is performed on the emergency rescue voice clip of the set elderly patient obtained through the voice collection device connected to the Internet of Things, and when the character string obtained by analysis matches the semantics of a certain emergency rescue type, a clear rescue voice signal is issued, and when the character string obtained by analysis does not match the semantics of any emergency rescue type, a fuzzy rescue voice signal is issued, including: the semantics of a certain emergency rescue type include synonymous character strings and near-synonymous character strings of the corresponding type name;

[0077] Wherein, while sending a clear signal for the emergency voice, the uniform division of the emergency voice segment on the time axis is temporarily suspended;

[0078] And wherein, while sending out a clear distress voice signal, a certain emergency distress type that matches the semantics of the character string obtained through analysis is output as the current emergency distress type.

[0079] Example 2

[0080] Figure 3 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 2 of the present invention.

[0081] like Figure 3 As shown, Figure 2 Different from the embodiment in, before step S22, the method further includes:

[0082] Step S31: using a network configuration interface to respectively perform network parameter configuration operations on a voice acquisition device and a plurality of physiological parameter sensors connected to the Internet of Things;

[0083] For example, using a network configuration interface to respectively perform network parameter configuration operations on a voice collection device and a plurality of physiological parameter sensors connected to the Internet of Things includes: the network configuration interface is based on a time division duplex communication mechanism or a frequency division duplex communication mechanism.

[0084] Example 3

[0085] Figure 4 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 3 of the present invention.

[0086] like Figure 4 As shown, Figure 2 Different from the embodiment in, after step S25, the method further includes:

[0087] Step S41: receiving the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient, and wirelessly sending the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient to the patient management server of the nearest nurse workstation;

[0088] Specifically, receiving the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient, and wirelessly sending the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient to the patient management server of the nearest nurse workstation includes: the patient management server of the nearest nurse workstation is a big data server, a cloud computing server or a blockchain server.

[0089] Example 4

[0090] Figure 5 The present invention is a flowchart of the steps of an Internet of Things-based voice emergency rescue method for the elderly according to Embodiment 4 of the present invention.

[0091] like Figure 5 As shown, Figure 2 Different from the embodiment in, after step S25, the method further includes:

[0092] Step S51: receiving the emergency help type code value corresponding to the emergency help voice clip of the set elderly patient, and displaying the corresponding type name of the emergency help type corresponding to the received emergency help type code value in real time at the bedside control panel of the set elderly patient;

[0093] For example, receiving the emergency help type code value corresponding to the emergency help voice clip of the set elderly patient, and displaying the corresponding type name of the emergency help type corresponding to the received emergency help type code value in real time at the on-site control panel of the bed of the set elderly patient, including: the on-site control panel of the set elderly patient's bed includes multiple display screens, multiple buttons, a Bluetooth transmission interface and a main controller.

[0094] Next, various method embodiments of the present invention are described in detail.

[0095] In the voice emergency rescue method for the elderly based on the Internet of Things according to various method embodiments of the present invention:

[0096] In each learning performed on the deep neural network, the known emergency rescue type code value corresponding to a certain emergency rescue voice clip issued by the set elderly patient in the past is used as the single output data of the deep neural network, and the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued an emergency rescue voice clip in the past, various related information of the set elderly patient and the respective signal amplitudes corresponding to each time evenly spaced on the time axis corresponding to a certain emergency rescue voice clip issued in the past are used as input data of the deep neural network. The execution of this learning action includes: when the set elderly patient issued an emergency rescue voice clip in the past, using multiple physiological parameter sensors connected to the Internet of Things to respectively obtain the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient;

[0097] Specifically, in each learning execution of the deep neural network, the known emergency rescue type code value corresponding to a past emergency rescue voice segment issued by the set elderly patient is used as the single output data of the deep neural network, and the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued an emergency rescue voice segment in the past, various related information of the set elderly patient, and each signal amplitude corresponding to each evenly spaced moment on the time axis corresponding to a past emergency rescue voice segment are used as input data of the deep neural network. Executing this learning action also includes: selecting to use FPGA devices to complete the measurement and simulation of each learning execution of the deep neural network based on the numerical simulation mode.

[0098] And in the voice emergency rescue method for the elderly based on the Internet of Things according to each method embodiment of the present invention:

[0099] Obtain various related information of the set elderly patient, wherein the various related information of the set elderly patient is the disease type, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph, including: the historical average blood pressure of the set elderly patient is the average of the remaining multiple blood pressure values ​​obtained by removing the maximum and minimum values ​​of each blood pressure value collected by the set elderly patient at each time evenly distributed on the time axis of the day before the day using a blood pressure sensor connected to the Internet of Things;

[0100] Among them, obtaining various related information of the set elderly patient, the various related information of the set elderly patient is the type of disease, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph, and also includes: the historical average heart rate of the set elderly patient is the average of the remaining multiple heart rate values ​​obtained by removing the maximum and minimum values ​​of each heart rate value collected by the set elderly patient at each time evenly distributed on the time axis of the day before the day using a heart rate sensor connected to the Internet of Things;

[0101] And wherein, various related information of the set elderly patient is obtained, and the various related information of the set elderly patient is the disease type, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph. It also includes: the historical average blood oxygen saturation of the set elderly patient is the average of the remaining multiple blood oxygen saturation values ​​obtained by removing the maximum and minimum values ​​of each blood oxygen saturation value collected by the set elderly patient at each moment evenly distributed on the time axis of the day before the day using a blood oxygen saturation sensor connected to the Internet of Things.

[0102] Example 5

[0103] Figure 6 The figure is a schematic diagram of the structure of an elderly voice emergency rescue system based on the Internet of Things according to Example 5 of the present invention.

[0104] like Figure 6 As shown, the elderly voice emergency rescue system based on the Internet of Things includes a memory and one or more processors, the memory stores a computer program, and the computer program is configured to be executed by the one or more processors to complete the following steps:

[0105] Step S21: storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request;

[0106] For example, storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request respectively includes: the different emergency rescue type code values ​​are different binary code values;

[0107] Step S22: performing voice content analysis on the emergency help voice segment of the set elderly patient obtained through the voice collection device connected to the Internet of Things, and sending a clear help voice signal when the character string obtained through analysis matches the semantics of a certain emergency help type, and sending a vague help voice signal when the character string obtained through analysis does not match the semantics of any emergency help type;

[0108] Specifically, a voice content analysis is performed on a set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things, and when the character string obtained through analysis matches the semantics of a certain emergency rescue type, a clear distress voice signal is issued, and when the character string obtained through analysis does not match the semantics of any emergency rescue type, an ambiguous distress voice signal is issued, including: a mature voice recognition application on the market can be used to perform voice content analysis on the set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things;

[0109] Step S23: when receiving the fuzzy distress voice signal, evenly divide the emergency distress voice segment on the time axis to obtain the signal amplitudes corresponding to each time moment evenly spaced on the time axis;

[0110] For example, when receiving an ambiguous distress voice signal, evenly dividing the emergency distress voice segment on the time axis to obtain the signal amplitudes corresponding to each time evenly spaced on the time axis includes: obtaining a single signal amplitude of the emergency distress voice segment every 0.5 milliseconds on the time axis;

[0111] Step S24: obtaining various related information of the set elderly patient, wherein the various related information of the set elderly patient is the disease type, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph;

[0112] Specifically, obtaining various items of associated information of the set elderly patient, wherein the various items of associated information of the set elderly patient are the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph, including: a programmable logic device can be selected to obtain various items of associated information of the set elderly patient;

[0113] Step S25: obtaining the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from a plurality of physiological parameter sensors connected to the Internet of Things, using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various associated information of the set elderly patient and the signal amplitudes corresponding to each moment evenly spaced on the time axis, and mapping the corresponding emergency help type based on the intelligently predicted emergency help type code value to output as the current emergency help type;

[0114] For example, the MATLAB toolbox can be selected to complete the test and simulation of the data processing process of using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis;

[0115] The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the hospitalization duration of the set elderly patient;

[0116] For example, the number of times the deep neural network is learned is positively correlated with the number of types of different emergency rescue types, including: the greater the number of types of different emergency rescue types, the greater the number of times the deep neural network is learned;

[0117] And by way of example, the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the set length of hospitalization of the elderly patient, including: the set length of hospitalization of the elderly patient is 10 days, the number of hidden layers is 3, the set length of hospitalization of the elderly patient is 15 days, the number of hidden layers is 4, the set length of hospitalization of the elderly patient is 20 days, the number of hidden layers is 5, and so on;

[0118] The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the length of hospitalization of the set elderly patient. The deep neural network model includes: in each learning performed on the deep neural network, the emergency rescue type code value corresponding to a known emergency rescue voice segment issued by the set elderly patient in the past is used as the single output data of the deep neural network, the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued an emergency rescue voice segment in the past, various related information of the set elderly patient, and each signal amplitude corresponding to each time evenly spaced on the time axis corresponding to a past emergency rescue voice segment are used as input data of the deep neural network to perform this learning action;

[0119] Wherein, obtaining the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from multiple physiological parameter sensors connected to the Internet of Things includes: the multiple physiological parameter sensors connected to the Internet of Things are a blood pressure sensor, a heart rate sensor and a blood oxygen saturation sensor;

[0120] Among them, the voice content analysis is performed on the emergency rescue voice clip of the set elderly patient obtained through the voice collection device connected to the Internet of Things, and when the character string obtained by analysis matches the semantics of a certain emergency rescue type, a clear rescue voice signal is issued, and when the character string obtained by analysis does not match the semantics of any emergency rescue type, a fuzzy rescue voice signal is issued, including: the semantics of a certain emergency rescue type include synonymous character strings and near-synonymous character strings of the corresponding type name;

[0121] Wherein, while sending a clear signal for the emergency voice, the uniform division of the emergency voice segment on the time axis is temporarily suspended;

[0122] And wherein, while sending out a clear distress voice signal, a certain emergency distress type that matches the semantics of the character string obtained through analysis is output as the current emergency distress type;

[0123] like Figure 6 As shown, illustratively, M processors are provided, where M is a natural number greater than or equal to 1.

[0124] Example 6

[0125] Figure 7 The figure is a schematic diagram of the structure of an elderly voice emergency rescue system based on the Internet of Things according to Example 6 of the present invention.

[0126] like Figure 7 As shown, the elderly voice emergency rescue system based on the Internet of Things includes the following components:

[0127] A data storage device, used to store different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request;

[0128] For example, storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request respectively includes: the different emergency rescue type code values ​​are different binary code values;

[0129] A content judgment device is used to perform voice content analysis on a set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things, and when the character string obtained through analysis matches the semantics of a certain emergency rescue type, send a clear distress voice signal; and when the character string obtained through analysis does not match the semantics of any emergency rescue type, send an ambiguous distress voice signal;

[0130] Specifically, a voice content analysis is performed on a set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things, and when the character string obtained through analysis matches the semantics of a certain emergency rescue type, a clear distress voice signal is issued, and when the character string obtained through analysis does not match the semantics of any emergency rescue type, an ambiguous distress voice signal is issued, including: a mature voice recognition application on the market can be used to perform voice content analysis on the set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things;

[0131] A segmentation and collection device connected to the content judgment device is used to evenly segment the emergency rescue voice segment on the time axis when receiving the fuzzy distress voice signal, so as to obtain the amplitudes of the signals corresponding to the evenly spaced moments on the time axis;

[0132] For example, when receiving an ambiguous distress voice signal, evenly dividing the emergency distress voice segment on the time axis to obtain the signal amplitudes corresponding to each time evenly spaced on the time axis includes: obtaining a single signal amplitude of the emergency distress voice segment every 0.5 milliseconds on the time axis;

[0133] An information parsing device is used to obtain various related information of a set elderly patient, wherein the various related information of the set elderly patient is the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, a character string corresponding to a standard paragraph, and a standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph;

[0134] Specifically, obtaining various items of associated information of the set elderly patient, wherein the various items of associated information of the set elderly patient are the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph, including: a programmable logic device can be selected to obtain various items of associated information of the set elderly patient;

[0135] A request identification device is connected to the content judgment device, the segmentation and collection device, the information analysis device and the data storage device respectively, and is used to obtain the current blood pressure, current heart rate and current blood oxygen saturation of a set elderly patient from multiple physiological parameter sensors connected to the Internet of Things, and use a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient and the signal amplitudes corresponding to each moment evenly spaced on the time axis, and map the corresponding emergency help type based on the intelligently predicted emergency help type code value to output as the current emergency help type;

[0136] For example, the MATLAB toolbox can be selected to complete the test and simulation of the data processing process of using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis;

[0137] The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the hospitalization duration of the set elderly patient;

[0138] For example, the number of times the deep neural network is learned is positively correlated with the number of types of different emergency rescue types, including: the greater the number of types of different emergency rescue types, the greater the number of times the deep neural network is learned;

[0139] And by way of example, the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the set length of hospitalization of the elderly patient, including: the set length of hospitalization of the elderly patient is 10 days, the number of hidden layers is 3, the set length of hospitalization of the elderly patient is 15 days, the number of hidden layers is 4, the set length of hospitalization of the elderly patient is 20 days, the number of hidden layers is 5, and so on;

[0140] The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the length of hospitalization of the set elderly patient. The deep neural network model includes: in each learning performed on the deep neural network, the emergency rescue type code value corresponding to a known emergency rescue voice segment issued by the set elderly patient in the past is used as the single output data of the deep neural network, the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued an emergency rescue voice segment in the past, various related information of the set elderly patient, and each signal amplitude corresponding to each time evenly spaced on the time axis corresponding to a past emergency rescue voice segment are used as input data of the deep neural network to perform this learning action;

[0141] Wherein, obtaining the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from multiple physiological parameter sensors connected to the Internet of Things includes: the multiple physiological parameter sensors connected to the Internet of Things are a blood pressure sensor, a heart rate sensor and a blood oxygen saturation sensor;

[0142] Among them, the voice content analysis is performed on the emergency rescue voice clip of the set elderly patient obtained through the voice collection device connected to the Internet of Things, and when the character string obtained by analysis matches the semantics of a certain emergency rescue type, a clear rescue voice signal is issued, and when the character string obtained by analysis does not match the semantics of any emergency rescue type, a fuzzy rescue voice signal is issued, including: the semantics of a certain emergency rescue type include synonymous character strings and near-synonymous character strings of the corresponding type name;

[0143] Wherein, while sending a clear signal for the emergency voice, the uniform division of the emergency voice segment on the time axis is temporarily suspended;

[0144] And wherein, while sending out a clear distress voice signal, a certain emergency distress type that matches the semantics of the character string obtained through analysis is output as the current emergency distress type.

[0145] In addition, the present invention may also cite the following technical contents to highlight the significant technical progress of the present invention:

[0146] Using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient and various signal amplitudes corresponding to various moments evenly spaced on the time axis, and mapping the corresponding emergency help type based on the intelligently predicted emergency help type code value as the current emergency help type output includes: performing octal value conversion processing on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient and various signal amplitudes corresponding to various moments evenly spaced on the time axis, and then synchronously inputting them into the deep neural network model;

[0147] Among them, using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient and each signal amplitude corresponding to each moment evenly spaced on the time axis, and mapping the corresponding emergency help type based on the intelligently predicted emergency help type code value as the current emergency help type output also includes: running the deep neural network model to obtain the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient represented by an octal value output by the deep neural network model;

[0148] The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of types of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the hospitalization time of the set elderly patient, and further includes: using a content conversion function to represent the content conversion relationship in which the number of times the deep neural network has been learned is positively correlated with the number of types of different emergency rescue types;

[0149] For example, when a content conversion function is used to represent the content conversion relationship in which the number of times the deep neural network is learned is positively correlated with the number of types of different emergency rescue types, the number of types of different emergency rescue types is used as input information of the content conversion function, and the number of times the deep neural network is learned corresponding to the number of types of different emergency rescue types is used as output information of the content conversion function.

[0150] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the purpose of the present invention, all of which are within the protection of the present invention.

Claims

1. A voice emergency rescue method for the elderly based on the Internet of Things, characterized in that: The method comprises: Storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request; Perform voice content analysis on a set elderly patient's emergency rescue voice clip obtained through a voice collection device connected to the Internet of Things, and send a clear distress voice signal when the character string obtained through analysis matches the semantics of a certain emergency rescue type; and send an ambiguous distress voice signal when the character string obtained through analysis does not match the semantics of any emergency rescue type; When receiving the fuzzy distress voice signal, the emergency distress voice segment is evenly divided on the time axis to obtain the amplitudes of the signals corresponding to the evenly spaced moments on the time axis; Obtaining various items of associated information of a set elderly patient, wherein the various items of associated information of the set elderly patient are the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, a character string corresponding to a standard paragraph, and a standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph; The current blood pressure, current heart rate and current blood oxygen saturation of a set elderly patient are obtained from multiple physiological parameter sensors connected to the Internet of Things, and a deep neural network model is used to intelligently identify the emergency help type coding value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis, and the corresponding emergency help type is mapped based on the intelligently predicted emergency help type coding value to output as the current emergency help type.

2. The method for elderly people to call for emergency help via voice based on the Internet of Things as claimed in claim 1, characterized in that: The deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the hospitalization duration of the set elderly patient; Among them, the deep neural network model is a deep neural network that has completed multiple learnings, the number of times the deep neural network has been learned is positively correlated with the number of different emergency rescue types, and the deep neural network includes multiple hidden layers, a single output layer and a single input layer, and the number of hidden layers is proportional to the hospitalization duration of the set elderly patient. It includes: in each learning performed on the deep neural network, the known emergency rescue type code value corresponding to a past emergency rescue voice segment issued by the set elderly patient is used as the single output data of the deep neural network, the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued an emergency rescue voice segment in the past, various related information of the set elderly patient, and each signal amplitude corresponding to each evenly spaced moment on the time axis corresponding to a past emergency rescue voice segment is used as the input data of the deep neural network to perform this learning action.

3. The method for elderly people to call for emergency help via voice based on the Internet of Things as claimed in claim 2, characterized in that: Acquiring the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from multiple physiological parameter sensors connected to the Internet of Things includes: the multiple physiological parameter sensors connected to the Internet of Things are a blood pressure sensor, a heart rate sensor and a blood oxygen saturation sensor; Among them, the voice content analysis is performed on the emergency rescue voice clip of the set elderly patient obtained through the voice collection device connected to the Internet of Things, and when the character string obtained by analysis matches the semantics of a certain emergency rescue type, a clear rescue voice signal is issued, and when the character string obtained by analysis does not match the semantics of any emergency rescue type, a fuzzy rescue voice signal is issued, including: the semantics of a certain emergency rescue type include synonymous character strings and near-synonymous character strings of the corresponding type name; Wherein, while sending a clear signal for the emergency voice, the uniform division of the emergency voice segment on the time axis is temporarily suspended; When a clear distress voice signal is sent, a certain emergency distress type that matches the semantics of the character string obtained through analysis is output as the current emergency distress type.

4. The method for elderly people to call for emergency help via voice based on the Internet of Things as claimed in claim 3, characterized in that: Before performing voice content analysis on a set elderly patient's emergency rescue voice segment acquired through a voice collection device connected to the Internet of Things, sending a clear distress voice signal when the acquired character string matches the semantics of a certain emergency rescue type, and sending a vague distress voice signal when the acquired character string does not match the semantics of any emergency rescue type, the method further includes: The network configuration interface is used to perform network parameter configuration operations on a voice acquisition device and multiple physiological parameter sensors connected to the Internet of Things.

5. The method for elderly people to call for emergency help via voice based on the Internet of Things as claimed in claim 3, characterized in that: After obtaining the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from multiple physiological parameter sensors connected to the Internet of Things, using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient and each signal amplitude corresponding to each moment evenly spaced on the time axis, and mapping the corresponding emergency help type based on the intelligently predicted emergency help type code value to output as the current emergency help type, the method also includes: Receive the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient, and wirelessly send the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient to the patient management server of the nearest nurse workstation.

6. The method for elderly people to call for emergency help via voice based on the Internet of Things as claimed in claim 3, characterized in that: After obtaining the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient from multiple physiological parameter sensors connected to the Internet of Things, using a deep neural network model to intelligently identify the emergency help type code value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient and each signal amplitude corresponding to each moment evenly spaced on the time axis, and mapping the corresponding emergency help type based on the intelligently predicted emergency help type code value to output as the current emergency help type, the method also includes: Receive the emergency help type code value corresponding to the emergency help voice clip of the set elderly patient, and display the corresponding type name of the emergency help type corresponding to the received emergency help type code value in real time at the on-site control panel of the bed of the set elderly patient.

7. The method for elderly people to call for emergency help through voice based on the Internet of Things as described in any one of claims 3 to 6, characterized in that: In each learning execution of the deep neural network, the known emergency rescue type code value corresponding to a past emergency rescue voice clip issued by the set elderly patient is used as the single output data of the deep neural network, and the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient obtained when the set elderly patient issued a past emergency rescue voice clip, various related information of the set elderly patient and the signal amplitudes corresponding to each time evenly spaced on the time axis corresponding to a past emergency rescue voice clip are used as input data of the deep neural network. Execution of this learning action includes: when the set elderly patient issued a past emergency rescue voice clip, multiple physiological parameter sensors connected to the Internet of Things are used to respectively obtain the past blood pressure, past heart rate and past blood oxygen saturation of the set elderly patient.

8. The method for elderly people to call for emergency help through voice based on the Internet of Things as described in any one of claims 3 to 6, characterized in that: Obtain various related information of the set elderly patient, wherein the various related information of the set elderly patient is the disease type, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph, including: the historical average blood pressure of the set elderly patient is the average of the remaining multiple blood pressure values ​​obtained by removing the maximum and minimum values ​​of each blood pressure value collected by the set elderly patient at each time evenly distributed on the time axis of the day before the day using a blood pressure sensor connected to the Internet of Things; Among them, obtaining various related information of the set elderly patient, the various related information of the set elderly patient is the type of disease, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph, and also includes: the historical average heart rate of the set elderly patient is the average of the remaining multiple heart rate values ​​obtained by removing the maximum and minimum values ​​of each heart rate value collected by the set elderly patient at each time evenly distributed on the time axis of the day before the day using a heart rate sensor connected to the Internet of Things; Among them, various related information of the set elderly patient is obtained, and the various related information of the set elderly patient is the disease type, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, character string corresponding to the standard paragraph, and the standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph. It also includes: the historical average blood oxygen saturation of the set elderly patient is the average of the remaining multiple blood oxygen saturation values ​​obtained by removing the maximum and minimum values ​​of each blood oxygen saturation value collected by the set elderly patient at each moment evenly distributed on the time axis of the day before the day using a blood oxygen saturation sensor connected to the Internet of Things.

9. An Internet of Things-based voice emergency rescue system for the elderly, characterized in that: The system comprises a memory and one or more processors, the memory storing a computer program, the computer program being configured to be executed by the one or more processors to perform the following steps: Storing different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request; Perform voice content analysis on a set elderly patient's emergency rescue voice clip obtained through a voice collection device connected to the Internet of Things, and send a clear distress voice signal when the character string obtained through analysis matches the semantics of a certain emergency rescue type; and send an ambiguous distress voice signal when the character string obtained through analysis does not match the semantics of any emergency rescue type; When receiving the fuzzy distress voice signal, the emergency distress voice segment is evenly divided on the time axis to obtain the amplitudes of the signals corresponding to the evenly spaced moments on the time axis; Obtaining various items of associated information of a set elderly patient, wherein the various items of associated information of the set elderly patient are the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, a character string corresponding to a standard paragraph, and a standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph; The current blood pressure, current heart rate and current blood oxygen saturation of a set elderly patient are obtained from multiple physiological parameter sensors connected to the Internet of Things, and a deep neural network model is used to intelligently identify the emergency help type coding value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis, and the corresponding emergency help type is mapped based on the intelligently predicted emergency help type coding value to output as the current emergency help type.

10. An elderly voice emergency rescue system based on the Internet of Things, characterized in that: The system comprises: A data storage device, used to store different emergency rescue type code values ​​corresponding to different emergency rescue types including bleeding control request, respiratory obstruction clearance, vital sign monitoring request, emergency defibrillation request, drug call request and emergency transport request; A content judgment device is used to perform voice content analysis on a set elderly patient's emergency rescue voice segment obtained through a voice collection device connected to the Internet of Things, and when the character string obtained through analysis matches the semantics of a certain emergency rescue type, send a clear distress voice signal; and when the character string obtained through analysis does not match the semantics of any emergency rescue type, send an ambiguous distress voice signal; A segmentation and collection device connected to the content judgment device is used to evenly segment the emergency rescue voice segment on the time axis when receiving the fuzzy distress voice signal, so as to obtain the amplitudes of the signals corresponding to the evenly spaced moments on the time axis; An information parsing device is used to obtain various related information of a set elderly patient, wherein the various related information of the set elderly patient is the type of illness, age, historical average blood pressure, historical average heart rate, historical average blood oxygen saturation, a character string corresponding to a standard paragraph, and a standard paragraph voice segment obtained by the set elderly patient reading the standard paragraph; A request identification device is respectively connected to the content judgment device, the segmentation and acquisition device, the information analysis device and the data storage device, and is used to obtain the current blood pressure, current heart rate and current blood oxygen saturation of a set elderly patient from multiple physiological parameter sensors connected to the Internet of Things, and use a deep neural network model to intelligently identify the emergency help type coding value corresponding to the emergency help voice segment of the set elderly patient based on the current blood pressure, current heart rate and current blood oxygen saturation of the set elderly patient, various related information of the set elderly patient, and each signal amplitude corresponding to each moment evenly spaced on the time axis, and map the corresponding emergency help type based on the intelligently predicted emergency help type coding value to output as the current emergency help type.

Images