Sage Journals: Discover world-class research

Abstract

With the acceleration of global aging, the service of the aged has been widely concerned by the whole society. The traditional service mode has been challenged, and it is difficult to meet the needs of the aged. In order to provide better services for the aged, the construction of emotional intelligent service based on Internet of things is proposed in this article. First, the framework is introduced, and the architecture of the whole system is given. Second, the key technologies of emotional intelligence and Internet of things are described, which mainly include affective computing and data collection of Internet of things. Third, the implementation of the software and hardware is given, and the core functional modules are analyzed. As a whole, there are two main contributions of this article. One is to give the framework of emotional intelligence service system for the aged based on Internet of things, and the other is to analyze the emotional state and the physical data of the aged so that the early-warning and intelligent services can be provided for them.

Keywords

Emotional intelligent service the aged Internet of things affective computing

Introduction

In recent years, the problem of the aged services has become an important issue in the world. It was reported that the population over the age of 65 years will reach one-fifth by 2050,¹ and the proportion of the people aged 65 years in many countries will exceed 25% compared with those aged 20–64 years by 2030, such as the United States is 36.6%, Germany is 48.6%, Japan is 57.1%, Korea is 40.7%, Singapore is 39.2%, and China is 27.7%.² Therefore, the demand for the service market for the aged is huge in the future. Service providers who perceive the physical condition of the aged and analyze the behavior of the aged can provide effective intelligent services, which is difficult to achieve with traditional technologies.³ Therefore, how to provide intelligent services for the aged has always been the focus of social attention.

With the application of cloud computing, Internet of things (IoT), and artificial intelligence technology in various fields, the aged services based on these technologies is also gradually developing. More and more countries, enterprises, and scholars try to apply the above technologies to the aged services. The United States is the first country to enter the aging society, and it is also the leading country of IoT technology. As the way of residence of the United States is different from that of China, they are not in the form of community, but mainly in the form of single houses. So, their way is home-based care for the aged, and radio-frequency identification (RFID) technology is widely used in the daily life of the aged. For example, some electronic labels are embedded in the daily life of the aged or in the daily wear of the aged, and some health indicators of the body are monitored in real time.⁴ Japan was the first to use the IoT technology in the field of life. In the 1990s, Japan used chip technology to facilitate home life in various remote controls, such as Wireless TV remote control, wireless dimmer and so on.⁵ In China, IoT technology is mainly used for home care after 2009. For example, the smart pension project launched in Nanjing city in 2010; the project served the aged by installing electronic chips in the floor of the aged, installing a temperature sensor on a kitchen stove, installing a wireless smoke detector in the kitchen, and so on.⁶

So far, many scholars have studied care for the aged such as home old-age service network center,⁷ intelligent medical technology for the aged,⁸ intelligent home system for the aged,⁹ and so on. For example, Ge and Wu¹⁰ studied the application of IoT technology in home care service and described the role of intelligent integrated service platform for the aged and service demand providers, and they point out the shortcomings of the application of the IoT in the home for the aged in China, which mainly shows that the coverage of the items in the home care service of the IoT is less and the level of information is low. Marques and Rui¹¹ built an environmental monitoring system and life support system based on the IoT to improve the health and comfort of the aged, and the system allows data collection through web access and mobile applications to support medical diagnosis. On the basis of traditional old-age and institutional pension in China, Qu et al.¹² put forward a kind of intelligent old-age community system based on the IoT, which includes intelligent old-age care subsystem, intelligent pension entertainment subsystem, intelligent old-age restaurant subsystem, intelligent pension shopping subsystem, and so on. In general, most of the previous studies are focused on the construction of the aged service system based on the IoT. There are few studies on the aged emotion and data analysis, and then, the emotional intelligence of the aged based on the IoT, which is just the content of our study.

This article is organized as follows. In section “Requirement analysis and framework,” the framework of emotional intelligent service for the aged is introduced. In section “The key technologies,” it describes the key technologies of emotional intelligence and IoT. The implementation of the software and hardware of emotional intelligence service is given in section “The implementation of the system.” Section “Conclusion” gives the summary of this article.

Requirement analysis and framework

Requirement analysis

It was well known that the lack of medical resources for the aged has existed for a long time in today’s society. Especially in China, it is estimated that the number of the aged over 60 years will increase to about 255 million by the end of 2020, and it will account for 17.8% of the total population.¹³ Statistics data from Shanghai shows that the amount of the aged over 60 years in Shanghai is 4.836 million by the end of 2017, which accounts for 33.2% of the household register population. There are 703 pension institutions with a total number of 1,404,000 beds in Shanghai, and the number of beds in pension institutions is less than 3% of the population of the aged.¹⁴ Therefore, how to provide efficient pension services is particularly important.

For a long time, the traditional family support has been undertaking the main function of providing for the aged. It is the main way to provide informal care, such as spouses, children, and other family members or other relatives. This kind of old-age mode mainly has the following problems: first, the caretakers are the children or others. Most of them lack professional nursing knowledge and cannot provide professional and comprehensive care. Second, the government’s policy and financial support for home care are insufficient, and it lacks necessary and comprehensive home care support. Third, the home-based care for the aged is mostly based on family units, which fail to achieve resource sharing, and once the problem occurs, it is easy to see the phenomenon of illness and emergency, and greatly increase the cost of human resources. In addition, it often happens that a couple needs to look after many aged, which makes them feel physically and emotionally drained.

In order to solve the above problems, the solution of emotional intelligence service for the aged has been put forward in our study, which is built on the IoT, including the functions of body monitoring service, location service, fall detection, call service, emotion judgment for the aged, and so on. The main users of the system include the aged, health care provider, and the family members of the aged. The requirements of the system are analyzed as follows.

The requirements of the aged

According to Maslow hierarchy of needs theory, human needs can be divided five types from low to high, which is physiological needs, security needs, social needs, need to be respected demand, and self-realization needs. For the aged, the needs of the first three tiers are even more important. In general, the aged must first ensure their own safety and physical health, ensure the normal physical signs, and enhance the emotional communication with their families based on the above situation. Considering the physical function and responsiveness of the aged, the interface and function of the software should be as simple as possible and avoid manual input. Most of them can be done by clicking.

The requirements of the health care provider

As an important part of the aged pension system, health care providers are the key members of the system. They can provide not only physical help for the aged but also provide targeted advice and help to the aged from lifestyle, diet, and reduce the probability of sudden illness in the aged from the source. The main information concerned by health care providers includes the vital signs, habits, and the disease history of the aged. For example, the heart, respiration, body temperature, and blood pressure of the aged.

The requirements of the family members

In today’s competitive society, young people are mostly busy working and living, which makes them have little time and energy to take care of the aged. Therefore, from the point of view of family members, it is necessary to understand the situation of the aged by a simple and quick way and without the need of professional medical knowledge, and it can help the aged to enjoy their old age better under the guidance and assistance of health care provider.

The requirements of system function

According to the survey, emotional intelligence service system should have the following functions.

Simple and rapid acquisition of vital signs for the aged.

Health care provider can use the system software to monitor the data manually.

Health care provider can analyze the current situation of the aged and put forward appropriate information for the aged.

It can send alarm information to health care provider when abnormal data are found in monitoring.

Family members can check the physical condition of the aged and endowment advice information through the Internet anytime and anywhere.

In addition, the software should be highly open for future maintenance, reuse, and development.

Framework

The system is composed of hardware and software, which includes data collection layer, intelligent home care service platform, and the system of application layer (e.g. hospital management information system). On the basis of considering the needs of users and referring to the previous intelligent service system, the system combines the advanced sensor technology, Internet technology, emotional computing technology, and intelligent terminal technology. The framework of the system is shown as in Figure 1.

Figure 1.

The framework of the system.

The hardware system

Hardware system includes sensors of IoT, intelligent terminal, and integrated health machine. The sensors collect the temperature, humidity, and video data from the home environment, and send it to the application server regularly. Intelligent terminals (such as smart watches) can measure the blood pressure, heart rate, and body temperature of the aged, and realize the function of location and emergency call. The integrated RFID function can be used for user identification. The integrated health machine is installed in a fixed area of the community. After reading user information through RFID card, it collected health data such as heart rate, blood glucose, blood pressure, blood oxygen, urine routine, body temperature, body weight, uric acid, and total cholesterol and then uploaded to the application server.

Sensors transmit data through WiFi or 2G/3G/4G networks to the servers via the middleware. The server receives data and stores it. When customer service receives the call from the aged, they will respond to the needs of the aged immediately and distribute the tasks. The management center can view the data statistics and details through the service board, and the aged family’s member can check the health status and environment data of the aged through the mobile App. When an emergency call or data exception occurs, the management system will inform the aged and their family’s member in the first time through App messages and SMS.

The software system

The software system includes IoT subsystem, home management subsystem, emotional service management subsystem, and health records subsystem. The function of the software system is shown as in Figure 2.

Figure 2.

The function of the software system.

The subsystem of IoT includes sensor management, data receiving middleware, statistical data analysis, and other functions. Health, environment, location, and other information are collected by sensor devices in the IoT, and are uploaded to the server regularly. After receiving the data, the system analyzes and takes the data. If it is abnormal data, it should take alert and send alarm information.

The subsystem of home management includes home management and information management for the aged. Home management includes information on home location, facilities, environment, and so on. The management of the aged includes basic information, health records, and so on.

The subsystem of service management includes call management, cooperator management, volunteers’ management, and service management. The service centers answer and track the incoming calls, WeChat, QQ, email, and other messages. The cooperator management realizes the business qualification recognition, service quality tracking, and other functions. The service management consists of tracking and evaluation of service orders generated, assigned, and served.

The subsystem of health records includes health file, third-party application interface, data statistical analysis, and other functions.

The role of user

There are four user identities in the system, which is the aged, their families, cooperators, and administrators. The aged can call the customer service through the one-button call function which integrated in the smart watch. In addition, they can also seek service and support through websites, App, and WeChat.

The family members can select or customize services for the aged through websites, App and WeChat, and conduct service evaluation.

Cooperators receive orders from the platform, provide services, and accept the supervision of the platform. After receiving the task, the staff of the service organization will carry out the service according to the agreed time. After the completion of the service, they will upload photos, and the aged or their family members will evaluate the service. Take home care as an example, the process for service is shown in Figure 3.

Figure 3.

The process for home care service.

The key technologies

Protocol conversion

Taking into account the aged particularity, we adopted wireless transmission method to collect their data based on IoT. The IoT connects anything to the Internet through information sensing devices such as RFID, sensors, GPS, and laser scanners. The exchange of information between devices is through the Zigbee, Wave, Xmesh, and Smart Mesh/TSMP protocol. The protocols used in the Internet are mainly TCP/IP. In order to connect to the Internet and read or write information, it is necessary to consider the protocol conversion.

Zigbee is a set of 802.15.4 protocols approved by IEEE. It is a technical standard for wireless sensor networks, security, and application software. Because of the advantages of low cost, high reliability, low power consumption, and strong environmental adaptability, Zigbee protocol is widely used. They communicate with the Internet through a Zigbee–TCP/IP gateway for protocol conversion.

The conversion between Zigbee protocol and TCP/IP is realized through two-way conversion of packets. The protocol transformation principle is shown as in Figure 4.

Figure 4.

The conversion of Zigbee and Ethernet Protocol.

Affective computing

Emotions need to be categorized before performing affective computing.¹⁵ Up to now, scholars have divided human emotions into discrete emotions and continuous emotions.^16,17 Among them, discrete emotions include such as happy, anger, sad, and fear. Continuous emotion refers to mapping emotions into space and dividing emotions based on spatial models.

PAD emotional model

PAD emotional model is based on previous psychological research results, and it is a dimension measurement model that can express multiple emotions.¹⁸ The PAD emotion model holds that human emotions can be expressed in three dimensions, namely, pleasure, activation, and dominance. In the PAD emotion model, the values of the three dimensions (P, A, and D) are from −1 to 1. Compared with traditional emotion description methods, each emotion in PAD emotion model can correspond to the location of PAD space. When the PAD score is standardized, it can express emotion with the unique three-dimensional coordinates and can describe the differences of various emotions in more detail. And, it can be used to describe and compare complex and subtle emotions. For example, Zhang et al.¹⁹ used the PAD value to study on facial expression emotion, shown in Table 1.

Table 1.

Sample of PAD for facial expression emotion.

State	Value
State	P	A	D
Angry	−0.59	0.08	0.47
Disappointment	−0.59	−0.01	0.40
Fear	−0.08	0.18	−0.39
Happiness	0.63	0.40	0.29
Calm	0.03	−0.04	−0.07
Sadness	−0.28	−0.12	−0.37
Surprise	0.41	0.55	0.19

When a person is in a certain emotional state, his facial features will be different. Generally speaking, when a person is happy, he may open his mouth wide, his lips may be pulled back, and his eyebrows may be slightly bent. On the contrary, when a person is in pain, he tends to pull down the corners of his mouth, raise his upper eyelids, and wrinkle his eyebrows easily. Considering the characteristics of facial expressions in the aged, our study revised the coordinates of the facial features according to empirical data, which include the calculation of the coordinates of the mouth, eyes, and eyebrows. At the same time, speech is also applied to the emotional state recognition of the aged. And, sample of some applications and some parameters is shown in Table 2.

Table 2.

Sample of some applications and some parameters.

Application	Parameters	Description
Speech emotion recognition	Short-time energy, pitch and short time average zero crossing rate, 12-order Mel frequency cepstral coefficient, and so on.	If an old man feels a little afraid, some physiological phenomena will appear to him.For example, speaking speed is fast, or average pitch is higher, or the intensity of sound is low.
Facial expression emotion recognition	The position of the mouth, noise, eyes, eyebrows, apex nasi and so on.	If an old man feels a little afraid, some facial expression will appear to him.For example, when the old people feel fear, their eyebrows tend to rise and straighten.

Emotion recognition by deep convolution neural network

The neural network face recognition technology obtains the recognition model through the training of a large number of sample images and then identifies it through the recognition model. Neural network face recognition technology does not require manual selection of features and can be learned in the sample training process. Face recognition based on deep learning is a research hotspot in recent years.

The deep network architecture used in this article is a model of the convolution neural network (CNN) superimposed on multiple hidden layers.²⁰ It is obtained by adding the pool layer and the convolution layer to the perceptual machine. Each volume layer contains multiple feature maps, as shown in Figure 5.

Figure 5.

Mapping relation of convolution layer.

The input data of the convolution layer come from the network input layer or the implied pooling layer. In the convolutional layer, the neurons of each feature map are connected to the local accepted domain of the previous layer. Moreover, the neuron perception weights in the same feature map are the same value.

In CNN networks, the iterative updating of training parameters can be calculated on the basis of residual values of neurons. In the input layer, the connection weights of the i node neurons of the feature graph need to be summed up by calculating the contribution of each neuron to the weight in j. The calculation formula is as follows

$\frac{\partial J (W, b)}{\partial W_{ij}^{l}} = \sum_{u, v} {(P_{i}^{l})}_{uv} {(δ_{j}^{l + 1})}_{uv}$ (1)

where $(P_{i}^{l})_{uv}$ represents the collection of [u, v], and the corresponding neurons in the field of perception. $(δ_{j}^{l + 1})_{uv}$ refers to the corresponding residual characteristic map of the j characteristic map in the convolution layer.

For the value of bias b parameter, we only need to sum up the residual values of all the neurons in the volume layer. The calculation formula is as follows

$\frac{\partial J (W, b)}{\partial b_{i}^{l}} = \sum_{u, v} {(δ_{i}^{l + 1})}_{uv}$ (2)

Then, the parameters of deep CNN can be updated according to equations (1) and (2)

$\frac{\partial J (W, b)}{\partial W_{ij}^{l}} = \frac{\partial J (W, b)}{\partial X_{i}^{l + 1}} \cdot \frac{\partial X_{i}^{l + 1}}{\partial W_{ij}^{l}} = δ_{i}^{l + 1} a_{i}^{l}$ (3)

$\frac{\partial J (W, b)}{\partial b_{i}^{l}} = \frac{\partial J (W, b)}{\partial X_{i}^{l + 1}} \cdot \frac{\partial X_{i}^{l + 1}}{\partial b_{i}^{l}} = δ_{i}^{l + 1}$ (4)

Intelligent warning service

The aged intelligent warning service mainly includes the physical location and physiological parameters, such as falls, blood pressure, electrocardiogram, body temperature, and other abnormal conditions. Blood pressure is the side pressure of the blood vessel wall of the blood unit area. It mainly refers to the arterial blood pressure, that is, when the blood vessels dilate, the blood pressure drops and the blood pressure rises when the blood vessel is contracted. In each cardiac cycle, the heart is excited by pacemakers, atria, and ventricles, and changes in bioelectricity are called electrocardiography. Body temperature is an important characteristic and index of the body’s life state. Usually, the measurement of the body temperature characteristics is mostly done by a temperature sensor with the negative temperature coefficient.

In order to do well the intelligent alarm service for the aged, some data mining methods are used, such as classification, clustering, and association rules analysis. Among them, association rules analysis can be considered that if a transaction includes A, then the transaction may include B at the same time. The value of support can be described as follows

$Suppor t_{(A \to B)} = \frac{S_{(A)} \cap S_{(B)}}{S_{all}}$ (5)

The degree of confidence about emotional intelligent service for the aged can be described as follows

$Confidenc e_{(A \to B)} = \frac{S_{(A)} \cap S_{(B)}}{S_{(A)}}$ (6)

When some phenomena are discovered and their rules are excavated, association rules are formed. Emotional intelligence services can be done based on these rules. For example, the phenomena of “In a calm state, the heart rate of the aged is higher than 100 times or less than 60 times” is related to “Arrhythmia.” And, the other phenomena is that “in a calm state, The SBP (Systolic Blood Pressure) is between 130-139 mmHg or the DBP (Diastolic Blood Pressure) is between 90-109 mmHg,” which is related to “The warning level is medium.”

The implementation of the system

The intelligent home care system is developed by Microsoft Visual Studio 2015 development tools, c# language, Microsoft SQL Server 2012, and Android mobile Studio 3.1.2 development tools. The intelligent home care system includes the sensing layer, network layer, and application layer. The technical architecture is shown in Table 3.

Table 3.

Technical architecture design.

Application layer	Website, App, Smart device, Call center, Client of management system
Network layer	Cloud computing, Artificial intelligence, Database and data mining, Mobile network, Internet, WiFi, Satellite communications.
Sensing layer	Sensor, QR code, RFID, video collection equipment, distributed information processing, and embedded technology.

RFID: radio-frequency identification.

Sensing layer

The perception layer is the initial link of the intelligent home care system, including information collected by the aged and the home environment, such as temperature, humidity, health information, blood pressure, heart rate, and temperature. It is uploaded to the server through WiFi or 2G/3G/4G networks. It includes intelligent terminal, integrated health machine, and intelligent household monitoring.

Network layer

After acquiring information, the sensing layer relies on the network layer for transmission. At present, the main body of network layer is Internet, network management system, and computing platform. It also includes various heterogeneous networks and private networks. The network layer realizes the two-way transmission of data and control information through gateway, access network, and core network. In intelligent home care system, network layer includes 2G/3G/4G mobile network, wireless network, Ethernet, fiber network, and so on.

Application layer

The application layer is the interface of IoT and users. It uses information technology and visualization technology to make quick feedback to the demands of home care service. It connects the aged and service system. Meanwhile, it realizes the connection between demand and supply. The presentation form of application layer includes website, management system, App, micro website, WeChat, and so on. It includes website, IoT device management, home management, service management, health records management, and so on.

The core of intelligent home care system is website application. The aged and their families order services and products through the website. After receiving the order, the cooperator will arrange door-to-door service. The aged or their families can evaluate the service through the management platform or App. At the same time, the service quality is evaluated regularly by a third-party organization.

Health information, such as blood pressure, heart rate, and temperature of the aged, will be uploaded to the server. The system will perform data analysis. It will be informed via management platform, App, and SMS when abnormal data are found. The health data monitoring is shown in Figure 6.

Figure 6.

Health data monitoring.

It can be seen from Figure 6 that 5246 sensors based on IoT were used for terminal device services for the aged. Once there is an emergency to the aged, the system will give a yellow alarm in the corresponding location.

In general, the implementation of the emotional intelligent service for the aged system includes input and output modules. Among them, data collection is particularly important for the implementation of the services, and the relation of input data and output service is shown in Table 4.

Table 4.

The relation of input data and output service.

Input data	Output service
Data of environment perception such as temperature, humidity, noise, light intensity, and position information	Physical exercise service and early-warning service
Daily life data such as work and rest time, voice, facial expression, and body posture	Health care service and intelligent recommendation service
Data of indicators of body health, such as body temperature, blood pressure, and blood sugar	Health care service and early-warning service

Conclusion

With the rapid development of IoT, new intelligent devices are constantly emerging, which brings great changes to people’s daily life, especially in the service of the elderly. According to the service characteristics of the aged, we give a framework for the construction of an intelligent service system for the aged. In this framework, a unified family care service system has been established, which can provide community residents with life care, medical care, home service, spiritual comfort and other services. Because of the adoption of the IoT, the service has many advantages, such as real-time, fast, efficient, low cost, and security. In particular, the aged can make emergency calls through intelligent devices when an accident occurs. Moreover, through the comprehensive data analysis, we can find the hidden dangers of the aged and provide early warning of abnormal conditions for them, which is a very meaningful thing.

As a whole, there are two main contributions of this article. One is to give the framework of emotional intelligence service system for the aged based on IoT, and the other is to analyze the emotional state and the physical data of the aged so that the early-warning and intelligent services can be provided for them. In future research, more wearable devices²¹ can be used to collect the aged body data, and combining big data analysis²² to study the emotional intelligent service for the aged is the trend of the future.

Footnotes

Handling Editor: Muhammad Khurram Khan

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This research was supported,in part,by the International Cooperative Program of Shanghai Municipal Science and Technology Commission of China (grant no. 16550720500),Shanghai Higher Education Society (grant no. GJEL1851),Shanghai Philosophy and Social Sciences Plan (grant nos 2016BGL004 and 2018BGL023),Program of Shanghai Open University (grant no. HX1702),Major Projects of National Social Science Foundation of China (grant no. 16ZDA055),and Shanghai Young University Teachers Training Grant Program Project of Shanghai Municipal Education Commission (grant no. ZZSUIBE16026).

ORCID iD

Yonghui Dai

References

Kiani

Bayanzadeh

Tavallaee

et al . The Iranian population is graying: are we ready? Arch Iran Med 2010; 13: 333–339.

Chen

Research on key technologies of wisdom pension system based on Internet of Things. Wirel Inter Technol 2017; 5: 134–140.

Briconsouf

Newman

CR.

Context awareness in health care: a review. Int J Med Inform 2007; 76: 2–12.

Chernbumroong

Shuang

A practical multi-sensor activity recognition system for home-based care. Decis Support Syst 2014; 66(C): 61–70.

Heianza

The development strategies of Internet of Things in US, EU, Japan and Korea and the References to China. Sci Technol Progr Pol 2012; 29: e279–e285.

Chen

SQ.

Application and prospect of Internet of Things technology in the field of nursing care at home and abroad. Chin J Gerontol 2015; 5: 5349–5352.

Mohan

Gupta

Singh

et al . Study of depression in geriatric population: Old age home and community in Lucknow India. Int J Epidemiol 2015; 44: i97.

Narasimha

Agnisarman

Madathil

et al . An investigation of the usability issues of home-based video telemedicine systems with geriatric patients. Proc Hum Fac Ergonom Soc Ann Meet 2016; 60: 1804–1808.

Santoso

Redmond

SJ.

Indoor location-aware medical systems for smart homecare and telehealth monitoring: state-of-the-art. Physiol Meas 2015; 36: R53–R87.

10.

Application of Internet of Things technology in home care service in China. J Shanghai Univ Eng Sci 2016; 30: 267–271279.

11.

Marques

Rui

An indoor monitoring system for ambient assisted living based on Internet of Things architecture. Int J Environ Res Pub Heal 2016; 13: 1152.

12.

Jiao

Research of the wisdom community endowment system based on Internet of Things. J Shenyang Norm Univ 2017; 35: 93–97.

13.

Xinhua News Agency. Interpretation of the 13th five-year national plan for the development of the old age and the construction of the pension system, http://www.gov.cn/zhengce/2017-03/15/content_5177770.htm (accessed 15 March 2017).

14.

Shanghai Municipal People’s Government. Notice of “13th Five-Year plan” for the development of Shanghai’s old age career, http://www.shanghai.gov.cn/shanghai/download/gongkai/hff1685.pdf (accessed 30 September 2016).

15.

Dai

Han

Dai

et al . Emotion recognition and affective computing on vocal social media. Inform Manage 2015; 52: 777–788.

16.

Gong

Dai

et al . Emotion analysis of telephone complaints from customer based on affective computing. Comput Intell Neurosci 2015; 2015: 749326.

17.

Yang

Dai

WH.

Cost-sensitive learning for emotion robust speaker recognition. Sci World J 2014; 2014: 628516.

18.

Jia

Zhang

Meng

et al . Emotional audio-visual speech synthesis based on PAD. IEEE Trans Audio Speech Lang Process 2011; 19: 570–582.

19.

Zhang

Meng

et al . Facial expression synthesis using PAD emotional parameters for a Chinese expressive avatar. Lect Notes Comput Sci 2007; 4738: 24–35.

20.

Moon

Chang

Pan

SB.

A face recognition system based on convolution neural network using multiple distance face. Soft Comput 2017; 21: 4995–5002.

21.

Ragot

Martin

et al . Emotion recognition using physiological signals: laboratory vs. wearable sensors. In: International conference on applied human factors & ergonomics, Los Angeles, CA, 17–21 July 2017, pp.15–22. Cham: Springer.

22.

Baimbetov

Khalil

Steinbauer

et al . Using big data for emotionally intelligent mobile services through multi-modal emotion recognition. In: International conference on smart homes and health telematics, Geneva, 10–12 June 2015, pp.127–138. Cham: Springer.

Construction of emotional intelligent service system for the aged based on Internet of things

Abstract

Keywords

Introduction

Requirement analysis and framework

Requirement analysis

The requirements of the aged

The requirements of the health care provider

The requirements of the family members

The requirements of system function

Framework

The hardware system

The software system

The role of user

The key technologies

Protocol conversion

Affective computing

PAD emotional model

Emotion recognition by deep convolution neural network

Intelligent warning service

The implementation of the system

Sensing layer

Network layer

Application layer

Conclusion

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

References