Keywords

1 Introduction

1.1 Research Background and Motivation

Well-known journalist Fishman raised the major issues of population aging and a series of chain effects in his book “Shock of Gray” after visiting many cities and towns in various countries, aging not only affects the physical and mental safety and health of the elderly, but also affects the middle-aged, youth, and young children [1]. Therefore, society has gradually evolved many long-term care tragedies [2]. For example, many middle-aged people should be at the peak of their lives, but because they have elderly people to take care of, they must give up their careers and become the guardian of the elderly. However, the restricted lifestyle and time schedule have greatly restricted its development, and it will be more difficult to earn a living in the countryside. From an economic perspective, it is quite inefficient, but from an emotional and family perspective, it is priceless. In addition, there are many modern families who, because of their heavy burden of taking care of the elderly [3, 4], and because of the limited allocation of time resources, dare not to have the next generation, or sacrifice the growth education resources of the next generation.

Many countries are facing a severe population ageing problem [5,6,7]. The aging environment of the population, the trend of declining birthrate, the decline of family care function and the sense of loneliness in elderlies are impacting the lives of the elderlies. Therefore, in the environment of aging population, how to promote our society to achieve the Confucius philosophy “[8]… That makes the aged have the appropriate last years, those in their prime have the appropriate employment, and the young have the appropriate growth and development…” is an important issue of long-term care that we as a member of the aging environment cannot ignore.

Due to the rapid changes in social structure, countries around the world must develop long-term strategic planning and resource reorganization related to the care of the elderly [9]. And the rapidly aging population has always been the Taiwan government’s biggest concern. With the rapid growth of the elderly population, the resulting long-term care needs and home care responsibilities will become heavier and heavier. In order to establish a long-term care system that meets the needs of the elderly and the physically and mentally disabled, the Taiwan Executive Yuan adopted the “Long-Term Care Ten Year Plan 2.0” in 2016. At the same time, the Taiwanese government has strongly used various IoT applications in the “Long-Term Care Ten Year Plan 2.0” and hopes to improve the long-term care system with the intervention of in-formation and communication technology [10, 11].

Even though there are many long-term care services at the national level it has been in implementation, there are still many restrictions on the recipients receiving service. Many elderly people in the healthy life stage are not eligible for service. However, standing in the concept of “prevention is better than cure”, these healthy elders needs to be cared for in order to prevent the heavy resources and manpower generated in the future stage of treatment. The demand for long-term care is not only to meet their physiological needs, but also for their psychological needs.

In addition, Taiwan is a highly democratic society [12], and care for the elderly living in various regions should not deviate from the fairness and equality of social welfare due to geographical location, resource allocation and sparse population conditions. Moreover, the right to life and health of the basic hu-man rights of the elderly living in rural areas should not be put at a disadvantage. The United Nations introduced “Proclamation on Aging” as early as 1991, revealing the five principles of “independence, participation, care, self-realization, and dignity” that older people should have [13]. The WHO pro-posed an “active ageing” policy framework in 2002, advocating the three aspects of health, participation and safety to improve the quality of life of the elderly [14]. The Ministry of Health and Welfare in Taiwan has established an active aging indicator framework and monitoring system since 2015 [15]. Aiming at important impact factors and key results, develop a data collection and intervention model that spans the health process, across levels, and across fields, and establishes a data warehouse and decision support system. Promote the successful and active aging of individuals and society, reduce cowardice, disability, dementia, care costs and premature death.

1.2 Research Purpose

In a highly democratic society, care for the elderly living in various regions should not deviate from the fairness and equality of social welfare because of geographical location, resource allocation, and sparse population conditions, let alone the basic human rights of rural elderly living at a disadvantage with the right to health. The United Nations introduced the “Proclamation on Aging” as early as 1991, revealing the five principles of “independence, participation, care, self-realization, and dignity” that older people should have. The World Health Organization (WHO) proposed an “active ageing” policy framework in 2002, advocating to improve the quality of life of the elderly from the three aspects of health, participation and safety [16]. The Ministry of Health and Welfare has established an active aging indicator framework and monitoring system since the Taiwan in 2015. In response to important impact factors and key results, it has developed a cross-health history, cross-level, and cross-domain data collection and intervention model, and established a data warehousing and decision support system. Promote the successful and active aging of individuals and society, reduce cowardice, disability, dementia, care costs and premature death.

Thus, this study mainly establishes an active aging AI community care ecosystem as the main task. With the success of this active aging AI community care ecosystem, it can help to deal with the problems faced by aging, effectively prevent and reduce the waste of healthcare resources at early stage. With the current state of aging in advanced countries, AI healthcare robots is a vast market to be developed. Therefore, this study hopes to use the standardization and information privacy protection for the elderly to care for huge amounts of health care data, and apply machine learning and statistical techniques, and deep learning to find out the hidden information of the elderly care, to find key decision points, and to establish alert and notification mechanism. Through this process, this study hopes to not only care for the physical and mental health of the elderly, but also to take care of the mental status of the elderly and relatives around the elderly, and to establish a blessed smart community care sharing model.

Every elder is a treasure with rich experience and it is the love of many family members and friends. How to protect the last mile of the dignity of the elders in the life stage and make them older is our major mission and responsibility. Therefore, based on the severe resource aging and resource issues faced by elder care mentioned above, this research will apply new-generation information technologies such as big data, deep learning, the Internet of Things, artificial intelligence and blockchain technology to bionic technology Robot dog as a carrier (considering the effects of the Valley of Terror theory, this study does not use robots as a carrier to provide services), to build a comprehensive and community-based AI community elder care program for the elders, which is close to the elders’real lives and needs Based on the researcher’s mission and responsibility, design a care plan for the elderly on the last mile of life. And in combination with the development of telecommunications technology to minimize the possible inequality between the urban and rural areas and the gap. Finally, through the intervention of technology, the life stories and experiences of the elders can be passed down and become the treasures of the community and society in the present and the future.

2 Artificial Intelligent Applications in Elderly Care

In recent years, the development of telematics has been quite rapid and matured and diversified [17]. For example, Japan is a best example of the use of next-generation telematics technology in elderly care. Artificial Intelligence (AI) is to make computers through a series of deep learning to make machines intelligent, and then used to solve real-life problems [18]. In many countries, robots have gradually been introduced as caregivers in response to manpower shortages. For example, Japan has been developing AI to support the elderly’s life since 2015, and has vigorously developed a robot that combines huge amounts of data and AI to help support future long-term service.

In Japan, robots are the main development direction of elderly care. At present, the robots on the market can be divided into three categories, namely: nursing support robots, self-assistance robots, and chat robots [19,20,21]. This study will provide comprehensive care services through the integrated solution of the nursing support robot and chat robot. But no matter which type, AI robots should have a human temperature, and currently the provision of human emotional connections and spiritual care is quite a missing link. The European Union has also developed some robots for elderly care programs, such as SocialRobot, ENRICHME… [22, 23]. Some researchers even shouted slogans to encourage everyone to seize the opportunity to develop smart elderly care robots [24].

Another study points out that the biggest criticism of such applications is that they may have deeper loneliness or fear of lack of humanity in robots [25]. AI robots should have a human temperature, and the current overall international development trend is currently quite lacking in the provision of human emotional connections and spiritual care. For a robot that is applied in daily life, it is very important to be able to grow with the changes in technology and needs. Because of this, research intends to use an enhanced deep learning method to allow AI robot to enhance themselves at any time. This type of growth way allows us to stay at the forefront of technology and serve humanity.

In 2019, Gartner proposed ten strategic technological trends with the core concept of “human-oriented smart space”, which is consistent with the direction of this project. The biggest needs of silver-haired people in China are physiological and safety needs [26], but in the process of accompanying the elderly, they often find that the elderly need more, but it is because of dignity and some factors that cannot truly convey the true needs And this often causes them to feel lonely and worthless, which is often the main reason for the elderly’s depression.

3 Active Aging

The World Health Organization (WHO) began promoting the concept of active aging in 2002 [27]. The main appeal of active aging is to enable the elderly to improve the quality of life in old age, and strive to improve the process of health, participation and safety. The vision of active aging requires the joint efforts of multiple parties, not just the health system alone.

In 2003, researchers in Taiwan began to study issues related to active aging, and eventually created an age-friendly city [28]. According to the “Global age-friendly cities: A guide” by WHO [29], there are eight major areas for friendly cities for the elderly: housing, transportation, telecommunications, accessible and safe public space, social participation, community and health services, work with volunteering, respect for the elderly and social integration.

Taiwan began to build the first age-friendly city program in Chiayi in 2010 [30], and counties and cities have subsequently joined the initiative. At the same time, it has also led to the development of the silver economy. The involvement of science and technology in the silver economy does not cost a lot of money, but to re-conceive and design the multi-level intervention of technology, environment, services and systems from an age-friendly manner.

The focus of active aging is on the concept of “prevention”. It is hoped that the prevention perspective can be used to affect the improvement of the health of the elderly and prevent the occurrence of disability. At the same time, in 2013, 1,672 community care points were deployed nationwide to promote the health promotion of the elderly [31]. It is hoped that this will make our elders not only live old but also live well.

4 Taiwan’s Current Long-Term Care Policy

At this stage, Taiwan’s long-term care policy mainly promotes long-term care 2.0, focusing on the realization of local aging and providing multiple care services for the elderly from family, community, and institutions. It is mainly to build community-based care-based communities to promote local elderly’s quality of life.

Long-term care 2.0 Establish community-integrated service centers (A), composite service centers (B), and alleys of long-term care station (C) in each township to provide care services with high accessibility, flexibility and convenience. At the same time, we also set up community-based dementia common care centers for dementia, provide related resources to build a safe community environment for dementia, and strengthen the prevention and delay of disability [32].

In addition, the National Health Agency also promotes related elderly friendly health environment and service plans [33], including: transformation of medical services, adjustment of medical care institutions, distribution of community services, and upgrading of counties and cities.

The transformation of medical services is mainly to build a high-quality chronic care network for the elderly; the adjustment of medical care institutions is to promote an elderly friendly care environment that promotes the dignity, participation and health of the elderly; the community service network is to build a community health promotion network that is active in the physical, mental and social functions of the elderly Road; county and city construction and upgrading is to promote an age-friendly city suitable for the active life of the elderly.

5 Active Aging AI Community Care Ecosystem Design Concept and Process

Based on the above-mentioned deficiencies in activating the lives of the elderly. This study considers the elders who have been in bed for a long time and have limited mobility. At present, they have been involved in the mechanism of nursing and caring staff and have been doing so for a long time. Therefore, this study first excludes these two types of elderlies, and only focuses on the health and sub-health elderlies that can take care of themselves. Conceptual care for seniors, prevention is better than cure. Therefore, this study hopes that this is a temperature-based application plan, which not only protects the elderly’s physical and safety, but also integrates the characteristics of the community into the elderly’s social partners in the old age and personalization. The integration of nursing care applications enables the elderly to actively age from the process of social and self-improvement. Finally, through the integrated application of chatbots, they interact with the elders and retain their life experiences and stories, which not only provides elders with full care, but also uses these valuable life materials as the treasure of humanity’s knowledge in the future.

In the past, there were many related researches on population aging using information technology. At the beginning, most of the development process was in the clinical medical and nursing level. For example, Li [34] proposed a bed monitoring system centered on the bed and not prejudicing the privacy of the elderly. Then, government officials have gradually adapted health information technology from the electronic medical records at the clinical end to the concept of “returning health information to patients” [35].

However, the elders are relatively weak in information ability and literacy, and there are obvious problems of unequal information for the above-mentioned health-related data. Therefore, the promotion of elders’ health is not only achieved through the introduction of information technology. To assist, we must also think about how to make the large amount of health information they generate can be effectively and properly applied based on professional health care background, not just partial fragmentation, but a perfect solution.

Many current researches only address fragmented current long-term care problems, but the development of information diffusion practices in the subject is quite limited due to the limitations of the elderly’s own conditions, such as: Chen [36] summarizes the leisure needs for active aging of middle-aged and elderly And health information needs to build active and aging leisure and health apps, but not all elderly people make good use of mobile phones and apps, and the development of health information assistants also lacks room for growth. In addition, there are also many home care design products for the elderly in the current environment. However, there are still many shortcomings in terms of functionality and meeting the needs of the elderly. In particular, the satisfaction and promotion of the social and self-realization of the elderly are still lacking.

Therefore, this plan designs an AI elder care solution using artificial intelligence technology combined with the new generation of information technology. Through professional and impartial third parties with health care knowledge and portable health care equipment, it provides care for the elderly in the community such as urban areas. Prevention is better than cure. I hope that through the design of this study, we will help rural elders who are far away from medical institutions. They will not sacrifice their own health and the right to equality in survival because of time and space constraints, so that the elderly is active.

In addition, it is hope that through the design of the AI community elderly care, the rural elderly can no longer be lonely, and the young people can leave the country and struggle more safely. Finally, human beings are group animals and should not be lonely or lack a stage to play because of their age and restrictions on their place of residence. Therefore, this study will design a mechanism for rural elders to meet social needs and respect needs. Toward self-realization, let the soul of the elderlies on the last mile of the village be rich.

The robot’s mission achievement is very good, but it supports logic. Although it lacks emotion, morality, self, and loyalty, we can give human logic to its logic. Just as in the movie Robot and Frank, Frank and Robot live and interact as normal [37]. Therefore, this research and development technology combines the knowledge of bionics to inject active temperature into the robot dog. Hamlin, Wynn, and Bloom [38] has used experimental methods to understand the distinction between good and bad infants at six and ten months. The study uses different design characters to show babies to show the positive events and animations with matching animations. Negative events prove that babies choose characters who are combined with positive events. It proves that babies have developed simple social ability and judgment ability in the early stage. Zoologists also point out that dogs understand our emotional changes by watching us [39].

Attila [40] points out that dogs understand the meaning of words with the left brain, and the meaning of intonation is distinguished with the right brain. Therefore, considering biological instincts, we will continue to use deep reinforcement learning (RL) to allow robot dogs to learn by observing the environment, and continue to self-grow at any time based on new data., And the data in the collection process for analysis and application in order to obtain maximum efficiency, efficiency and benefits.

Finally, the development of a chatbot model and the analysis results of the voice database to guide the elderly to share life experiences and stories, thereby retaining valuable life trajectory assets and becoming Taiwan’s huge database of elderly care with the concept of holistic care. In addition, this research focuses on system integration and data analysis and application, in order to minimize the risk of equipment instability and provide stable, perfect, and future-oriented care for the elderly. The AI community care ecosystem including following three main processes:

  1. 1.

    Design and build the prototype and basic capabilities of an elderly AI community care system and environment. This stage mainly integrates hardware, including smart bracelets, edge computing home butler cores, intuitive reflection cores, brain control cores, depth cameras, GPS receivers, array microphones, wireless network cards, servos, etc. Connected to each other.

  2. 2.

    Then, through the integrated environment of the previous stage, collect the physiological data of the community elders (heart rate, meal time, drinking water\work and rest\medication time, movement trajectory), image data (faces from different angles\expressions\relationships\contacts pictures), Audio file data (daily discourse, topical discussion, emotional classification, personal characteristics and preference attribute classification) are stored in the AI community care big data center established by this study.

  3. 3.

    By integrating these collected data, it can analyze personal care patterns, not only reminding the elderly to develop regular meals, rest, water and medication, and emotional and health precautions.

In this study, a four-legged robot is used as the carrier, that is, a robot dog. The reason why robot dogs are used instead of robots is based on the following advantages: superior load bearing and cross-terrain capabilities, more mature balance and movement functions, and the possibility of developing into pet shapes in the future to reduce the theory of valley of terror caused by robots. In addition, this study also combines wearable smart bracelets as companion components, mainly based on the following advantages, including: mature products with waterproof and dustproof capabilities, long standby time, years of experience, and diverse functional forms You can choose to transfer data directly with Android or iOS. This research device is composed of the core of the housekeeper, and with edge computing and blockchain technology, it can consume less power and could report at the first time. Furthermore, it can also avoid the worry of missing data or difficult backup. Once some terminals are broken, just replace the hardware and all the data will be stored in the blockchain. Because personal data has been de-personalized using SHA encryption, only you can find the correct backup from the blockchain, thereby reducing user concerns about information security and privacy protection of personal data (Fig. 1).

Fig. 1.
figure 1

Active aging AI community care ecosystem design concept

Full size image

This study mainly designed the following four major work items to complete the AI community care ecosystem:

  1. 1.

    AI elderly care robot dog and AI elderly care big data center planning, design and construction

The robot interpretation of the elderly’s behavior by analysis based on Google Atomic Visual Actions (AVA), this data set distinguishes human behavior into three categories (1) posture or movement performance during movement (2) person-to-item interaction behavior (3) person-to-person interaction behavior, and 96,000 individual-type actions are labeled, resulting in 210,000 Action tags. By extending this database and training AI to judge the behavior of the master, the advantage of this database is that it takes only three seconds to complete the interpretation.

  1. 2.

    Identification of the elderly

This method uses fast.ai as a template. The advantage of this method is that it is mature and uses python as the basic syntax to call various function libraries including C. In the first stage, I started training AI to identify the master with a deep learning course. First, I collected photos of the master from various angles to establish a training database and used this database as a template to start training AI. Use Google’s massive crawling text to collect many photos of people and things as a non-master feature database (Confusion matrix). At the beginning of the project, the AI was trained using a server computer. Load it into the electronic dog and test its accuracy to identify the owner. This study uses convolutional neural network image recognition as the basis. The visual images seen by AI dogs are first simplified by convolution operations and pooling methods to facilitate subsequent comparison and data transmission. The advantage of the convolution operation is that it retains the nature of 2D pictures but simplifies the amount of data and complexity and uses the GPU’s parallel operation capabilities to effectively complete the comparison of pictures and libraries.

  1. 3.

    Detection and application of classification methods for elder behavior and health status development with deep learning technology

We use deep learning technology to teach robotic dogs to classify elder behavior into moods, griefs, sadness, and other categories based on their facial expressions or physiological information. The basic and important skills for identifying facial expressions and emotions in the early stages of bionic interaction with human society. Each node is calculated from the weighted input values of multiple nodes in the previous layer. These weights can be adjusted to perform special image recognition tasks.

In order to train a deep neural network to recognize photos with happy faces, we show them happy pictures as raw data (image pixels) on the input layer. For example: when we know that the final result is hi, the network will recognize the pattern in the picture and adjust the node weights to minimize the errors of the category pictures judged as hi Every picture that shows favor will help optimize picture weight. With enough input information and training, accurately analyze and recognize patterns corresponding to hilarious expressions. Since the early elders have just joined the initial stage of the database establishment, the plan uses deep reinforcement learning. The deep reinforcement learning method can respond to the initial stage of our environment. Currently, the information is less complete. The robot dog acts based on the environment to maximize the expected benefits. In addition, under the environment’s feedback stimulus, the expectation of stimulus can be gradually formed gradually, resulting in the habitual problem of obtaining the maximum benefit.

In this study, the human action understanding data set AVA is used as an auxiliary database to establish an elderly behavior understanding database in Taiwan. With the help of the bracelet’s real-time tracking of health information, the health status is divided into three levels according to the emergency level.

  1. 4.

    Matchmaking for elder networks based on social network analysis and semantic analysis

Liu [41] reported that elders love social networking sites, more than 90% are LINE users, and nearly 60% also play Facebook. Using the connection between the robot dog and the mobile phone can act as a Bluetooth transmission terminal, presetting the basic voice activation function of several communication software. First save the record by streaming, develop the “Verbatim Tool” to convert speech into text, and gradually improve its conversion accuracy through deep learning. The word processing is turned into a dataset, and the natural language processing tool Google Natural Language API is used to analyze the positive and negative scores of each sentence, combined with the Chinese word segmentation system developed by the Central Research Institute of Taiwan to find out the characteristics of the elderly (keywords of interests or behaviors).

Use the semantic database to calculate topics or common behaviors that elders are interested in (such as singing, drawing, chatting, playing chess, etc.), and list keyword functions and analyze text or voice to collect data, and also analyze the types of attractions often visited by the elderly, the length of stay, the frequency and the time period through the trajectory data, and match up similar potential late-life partners. This function will calculate the score of each item with different score weights, so it may be necessary to establish a suitable formula after obtaining a more credible weight decibel through a questionnaire.

6 Conclusion

This study takes care of the physical and mental health of the elders through artificial intelligence technology. It also uses the judgment of physiological information to provide their concepts of health and wellness. The interaction between humans and dogs promotes the elders to carry out some healthy activities, thereby effectively shortening the elderly’s bedtime, improve physical health functions and functions, and make it a good social circle through the matching of social networks.

Finally, this research uses the blockchain structure as the material for the elderly, which not only allows the elderly to live happily and with dignity from the design and planning of this project, but also focuses on “people-oriented” for individualized care.

Kate Swaffer, chairman of the International Dementia Union, mentioned that although a patient with dementia is a patient, we should also allow him/her to participate in or engage in activities of interest, but his/her family or the institution takes the dominance of the demented, not the elderly [42]. The elder has his own life history and interests and can also express his feelings. Good care is based on him to help him develop his preferences. It is everything he decides, not standard arrangements. We should give basic respect to the human rights of the elderly and give him the right and opportunity to choose. Although the decline in function may affect the elderly’s decision-making, identification, language, sense of space, direction, etc., if we respect their human rights, we can still Provide different options to match their existing functions, giving him the right and opportunity to choose, not at the discretion of caregivers or institutions.

In this study, it is emphasized listening and respect. Let us understand the elders, accompany the elders, and listen to the elders. Through the practice of artificial intelligence technology and huge amounts of data, we understand and give the elders the real needs. Leaving a huge amount of valuable elderly assets in the field can also enable technology to help everyone more to leave more memories of the elders’ voices, words and pictures, so that the miss is no longer far away. Of course, while retaining the data, the elders need not worry about their own information privacy. The blockchain structure designed in this study will properly protect the elders ‘data and guard the elders’ human rights.