Keywords

1 Introduction

The overall interest in assistive and social robots has emerged during the last years. There are many different viewpoints and future plans using robots in health care services and diagnostics [1]. One major trend in elderly care sector is looking forward the development of robot technology. It is expected that number of aged people will be 1.5 billion by 2050, which is triple compared to 2010 level [2]. Increase in life expectancy is the positive thing, but it means that society has to take into account the emerging trend of cognitive disorders [3], and the demands for long-term care [4].

The European robotics strategy favors Internet-based tools in advancing health care and robotics. The strategy emphasizes that robots should be integrated into communication frameworks, and health monitoring systems should operate over the Internet [5]. Design of robots should be user-centered, but according to Tanaka et al. [6] the conventional design of robots still favors the technology oriented approaches, and neglects benefits of robots to users.

All the trends and challenges above has been the starting point for our aim to develop an approach for assessing the need for assistive and/or social robots. The objective is that the approach would assist end-users to evaluate if robots are needed, and which robot type would match best with users’ expectations. We have designed our approach to meet the future requirements of web-based evaluation tools. As far as we know, there are no web-based tools available and only few non-web based concepts for utilizing classification tools have been introduced before [6].

The nature of this study is explorative and it gives an insight into the position of designing and providing social and/or assistive robots. In addition to literature search, we have conducted questionnaire surveys, interviewed professionals as well as reviewed results of our previous robot research.

1.1 Concepts, Approaches, Factors and Technology for Defining the Robots

Studies regarding the acceptance of technology and related factors have been discussed in literature rather well. There are two main models for studying the acceptance of information technology; the Technology Acceptance Model (TAM) [7], where perceived usefulness and ease of use have been discussed, and the Unified Theory of Acceptance and Use of Technology Model (UTAUT) [8], which takes into account also a user’s age, gender, experience and voluntariness of use. According to Heerink et al. [9] TAM and UTAUT models neither take into account social aspects of interaction with robots nor are developed with the elderly users in mind. Therefore, Heerink et al. [9] have developed a new model and added perceived enjoyment, social presence, perceived sociability, trust, and perceived adaptivity issues into a model. The presented models above are useful for evaluating acceptance of technology but those do not assist to decide if robots are needed. The most advanced concept, compared to our aim, is presented by Tanaka et al. [6] where they present how to use the International Classification of Functioning, Disability, and Health (ICF) [10] for evaluating and designing assistive robots. The ICF classification offers the framework for assessing an individual’s level of functioning. Even if we did not find articles on web-based approaches, we want to present some articles, which discuss the relevant factors for implementing and accepting robots among elderly.

Flandorfer [11] has reported that sociodemographic factors have significant impacts on robot’s acceptance, but a user’s earlier experiences with technology mitigates an adoption process. It has been reported that factors like the living environment, a user’s physical and mental condition and cognitive skills should be taken into account [12]. In addition, religious and cultural backgrounds are reported to be significant factors in the acceptance of assistive robots [13]. Tong et al. [14] have stated that the current simulators and design tools are meant for process-oriented workflows not for designing the humanoid robots. Meng and Lee [15] have argued that the traditional industrial robot engineering approaches are inappropriate to tackle the problem areas such as user-friendliness. Wu et al. [16] have studied the elderly persons’ perceptions regarding the robot’s appearance and discussed the importance of social context in designing robots. Saborovski and Kollak [17] have argued that the needs of elderly people and their relatives must be taken into account in design process, but the care professionals’ experiences are overlooked in the technical development. de Graaf et al. [18] have presented that enjoyment and attractiveness of the robot are important factors in perception. Alaiad and Zhou [19] have studied the determinants of adoption of home healthcare robots, and reported that sociotechnical factors play important role. Michaud et al. [20] have presented the interdisciplinary and exploratory design methodology to develop an assistive mobile robot for homecare. Andrade et al. [21] have concluded that cost of robotics technology is still prohibitive issue, which limits the wide use of robots. Peine et al. [22] have studied the design processes targeted at older persons and proposed to consider them as active consumers of technology instead of passive recipients. Chibani et al. [23] have reported the recent challenges and future trends on ubiquitous robotics and argued that integration of web services and ambient intelligence technologies might offer fruitful options compared with the standalone robots. Obi et al. [24] have researched ICT promoting in Japan and suggested that more efforts are needed to exploit ICT to meet the challenges among ageing population. Linner et al. [25] have developed an approach which takes into account the association between the environment and robot technology and argued that integration of service robot systems into real world has been difficult because of the separate development of the human environment and robotics systems.

In sum, the articles reviewed revealed that there is no available a web-based approach, which may support the end-users’ and their stakeholders to assess the need for robots, and such an approach might be useful.

2 Methods

We inquired from 10 leading professionals who represented assistive technology units of the 10 biggest cities and hospital districts in Finland to evaluate; (a) would an approach to assess the need for personal robots be useful, (b) what kind of computer-based approaches hospital districts are using for assessing the need for robots for elderly if any, (c) what are the current procedures in selecting robots for elderly, and (d) is robots related advisory a part of their services. Those ten hospital districts cover about 90 % of Finnish population and cases, which are potential for assistive or social robots use. Professionals from 7 hospital districts replied, covering about 70 % of potential cases in Finland. Then we conducted a pilot survey among 33 wellbeing technology professionals (Median age 38, SD 7.7) and inquired them to evaluate the relevance of; (a) our approach pointed to elderly people, and (b) suggested factors in our approach. Of them, 11 had background from engineering, 15 from health and 7 from computing sciences. After that we tested validity of our approach by conducting the survey questionnaire among 64 elderly persons (Median age 70, SD 7.8), consisting both men and women. We carried out also a targeted literature search and evaluated 123 articles, which related approaches, concepts, tools, methods and variables in assessing the need for assistive or social robots, even if none of them directly presented web-based approaches.

The nature of this study was explorative and it emphasized descriptive statistics. The number of participants were too small for deeper statistical analysis such as logistic or multiple regression analyses.

2.1 Design of Approach

Based on our previous research, literature review and ICF, we created the list of seven aspects and 46 variables (Table 1), which might be relevant in assessing need for robots and for evaluating a type of robot (social and/or assistive). Aspect A concerns a user’s demography and profile. Aspect B concerns a user’s social relationships. Aspects C takes into account a user’s overall health and an aspect D takes into account a user’s functional capacity. Aspect E discusses a user’s skills and learning capacity and an aspect F takes into account a user’s possibilities to invest in robot. Aspect G concerns a robot’s deployment environment from a technology point of view. We classified the greatest part of the factors by a Likert scale from 1 to 5, excluding one social factor (B18) where we asked “Are you living alone or with someone?” and one health factor (C25) where we asked “What is your current need for rehabilitation?” where scales were from 1 to 3. In addition, an aspect A included an age, an occupation and an interest factors, which were not classified in ICF. Section A included also a question regarding a user’s interest area. The idea for inquiring the users’ interest areas was that robots would deliver relevant and interactive content to users.

Table 1. Topics in self-evaluation
Full size table

The basic idea behind an approach and an algorithm is that a person who has lively social life and good health and functionality does not need a robot. Correspondingly, a person who has a good health and functionality, but feels that social connections are poor, might need a social (companion) robot. For example, if a person state that he/she is able to meet a family member really often, he/she may not need a robot. If meetings are infrequently a person might need an assistive social robot (connected robot) which is able to provide him/her a connection to someone who is able to help if needed. The third option is a social robot (companion robot) regarding cases where a person is able to meet family members every so often but needs sometimes companion.

Further, a person who has limitations in health, functionality and social connections might need an assistive social robot, which is able to assist emotionally and/or physically. We have done two exceptions in the algorithm. If a person states that his/her perceived mental health is poor or very poor, a system does not recommend a robot at all. Corresponding case concerns a user’s skills, attitude and learning capacity. If a person states that his/her experience of technology, robots, and computers or willingness to learn new things is poor, a system does not recommend to use a robot.

3 Results

Table 2 presents evaluation of wellbeing technology professionals regarding some variables in our approach. They considered that aspects and variables were relevant overall but some adjustments would be needed. They considered that issues regarding users’ health, functional capacity and social networks were important. Major differences between distributions occurred regarding relevance to assess users’ former occupations, level of religiousness and possibility to invest in robots. Minor differences occurred regarding, e.g., acceptance of remote control.

Table 2. Evaluation of some factors from wellbeing professionals’ viewpoint
Full size table

The professionals commented and suggested the improvements to an approach. Their main criticism concerned elderly people’s ability to understand self-assessment questions being associated with technical issues. In addition, they said that some questions were too general and not able to reveal the real need for robots. Here are some examples of comments from them: “You are not asking directly what kind of robot a user might need?” “Could we ask directly how interested in are you to have a robot at home if a robot is able to entertain you?” “The questions should be more focused.” “Assessment tool is too long and some technical terms are difficult to understand.” “The elderly people that I interviewed were not able to understand some questions.” “Some examples how to use robots in daily activities would have clarified to elderly the need for robots.” “It would be better to assess the need for a robot together with a family member.” “Only few elderly people can afford to buy an expensive robot.” “I wouldn’t ask possibility to invest in a robot. It could insult someone.” “We should tell how much robots might cost; otherwise elderly people are not able evaluate possibilities to invest in.” “What are the cognitive skills?” “If an elderly person does not live alone, we should ask if his/her spouse would accept a robot.” “Asking only if they have fears is not adequate. We should ask what kind of fears do they have?” “The elderly people should have more information and knowledge about robots before assessing the need for.” Even if criticism was presented, the comments of professionals were positive overall. Many of them reported that assessments of need for a personal robot aroused elderly people’s interest towards robots. In addition, professionals commented that variables were relevant and questions were able to give an important insight into need for robots.

Pilot study among the elderly people.

We dichotomized test persons by their perceived need for a robot. Table 3 presents the most interesting variables and possible differences between the groups who stated to need robots, and those, who stated not to need robots. It is logical that those who need robots perceived that they would use robots more frequently than they who do not need robots. Both groups had rather good physical and mental health, but there were some differences in functional and cognitive capacity and ability to move body parts. Both groups had some experiences about Internet and applications, but experiences about robots were limited. However, attitudes towards robots were quite neutral, but financial resources to buy robots were poor.

Table 3. Differences in variables among two groups of elderly people
Full size table

Our approach reported that 58 % of respondents who perceived no need for robots just now, might be potential users of robot assistance anyway. Regarding another group, who perceived need for robots, the corresponding figure was 61 %, which means that in reality 39 % of them do not need robots. The biggest difference was found between type of robots needed. An approach emphasized more often assistive robots to them, who perceived to need robots, and to another group an approach could not make distinction between assistive and social robots.

Inquiry from 10 big cities.

Leading professionals from seven hospital districts replied and commented that there is no any web-based method for defining and selecting assistive and/or social robots for elderly. According to them, robots are still quite unfamiliar for elderly people and cases where customers are asking any kind of robots are still limited. However, professionals found interesting if there would be a web-based method, which is able to help both them and customers. For example, one health care professional stated that “We don’t have anything like that but we are definitely interested in if someone will develop a method”. Another professional commented that “I suppose that elderly care professionals are not ready for selecting an assistive robot for the customers because of lack of knowledge and tools”. The comments were positive overall and the professionals understood that the new era of assistive robots will break soon into market. However, they commented also that the way how they operate just now is quite traditional, and lot of introduction and training will be needed if robot technology will be embedded into health and elderly care sector.

4 Discussion

Even if there are some evidences that robots are able to activate and help the elderly people [26], we found that web-based tools for assessing the need for robots and for selecting robots are still missing even if modern ICT offers all the applications and tools for creating such a system.

Previous studies have reported that the robotics systems should be tailored according to users’ needs and characteristics [27] but the challenge seems to be that we do not have robust tools for evaluating customers’ needs. There are some good studies which discuss appearance, [16] acceptance, [16, 18] needs [27] and functions [28] of robots but there is still lack of methods which can be used into decision making and design processes of social and/or assistive robots.

Our approach introduces the new idea for defining the need for robots and selecting between social and assistive robots. Our approach followed, in theory, the TAM [7] and UTAUT [8] and Heerink’s models [9]. Those models are well established but do not define need for robots or type of robots. Our approach is conceptually tangential with the approach of Tanaka [6], which exploits ICF [10] and suggests, what functionalities robots should include. Our approach takes into account users’ social connections, deployment environment of robots and users’ interest areas, which are not evaluated in the Tanaka’s model. Our approach tries to figure the big picture of an elderly person’s life at the moment with respect to need for a robot.

The novelty of our approach is that it assesses the need for robots and makes distinction between social and assistive robots, being based on the physical, cognitive, social, motivational, learning and economic aspects. The suggested aspects and related factors are based on our previous research and discussions with health care, wellbeing and robotics professionals. In addition, the used variables can be tracked with ICF [10]. The concept has some limitations, for sure, but it attempts to fulfill the expectations and future trends. The professionals from 7 hospital districts in Finland stated that an approach for assessing the need for robots would be important. In addition, 33 wellbeing technology professionals found an approach interesting and capable to assist in evaluating the need for robots. The pilot-test among 64 elderly persons revealed that an approach was able to find distinctions between those who needed for robots and those who did not. The distinctions were minor because both group were homogenous and participants had good physical and mental health overall. The approach presented that some people who perceived to need robots, would not need those anyway. Vice versa some people who perceived not to need robots, would need robots to some extent.

The main limitation of our study was that the number of participants was small and more research is needed to validate the approach, variables, credits and those connections between. However, our aim was to introduce a conceptual approach and we will set the credits and connections in the later stage. Creating the credit system requires lot of algorithm and software development and it is one of our future targets. Another limitation concerned the suggested variables. We are not able to guarantee the validity and importance of some variables, and the number of variables needed. However, we present the comprehensive list of variables, which might be relevant.

We state that our approach might be an important step for advancing social robotics and it gives good insight for people who operate in elderly care or robotics sectors. The feedback from elderly care and technology professionals strengthened that the web-based system for defining the need for robots is still missing but very welcome.