3. Human-Robot Collaboration

There are around 50 million worldwide living with dementia. The number of people with dementia worldwide is increasing, while the number of related healthcare workers is going down. One solution to this would be using social robots to take over some tasks performed typically by human caretakers. Social robots are machines that aim to solve a particular role by following social behaviour. In prior work, they have been used for recreational activities, reminders, and encouraging exercise. 

Culture is a significant factor since the different regions have different attitudes towards robots. due to sociocultural factors. Countries like Japan or the Netherlands have a high acceptance rate, while the US and Italy are much less acceptance. Making changes to the robot depending on the country or population is an important factor in ensuring that it is seen as a legitimate tool for elderly care. Another important factor is exposure to prior advanced technologies like smart devices, computers or robots. This means that researchers are required to give more directions or that participants may not participate at all. Notably will this type of acceptance increase over time since the future elderly will have had much more exposure to technology than the current generation. At the moment can this limitation be overcome by performing longer studies, which research has shown leads to a higher acceptance rate from users. Interestingly is the usability of a robot that does group conversation much lower than that of a robot focusing on daily activities or elderly needs. 

Important future directions are user acceptance, perceived usefulness and perceived ease of use. An example of this would be regional accents for robots, informal word choice or friendlier voices. Utilizing a focus group would be an effective strategy to gauge the acceptance of the robots by the patients. 

Since dementia is a progressive illness the load of the caretakers increases over time. Easing the burden on the caretaker is therefore a vital field of research for social robots. It is therefore important to get the opinion feedback of caretakers, especially since they are responsible for the well-being and safety of people with dementia. This ensures that the design, development and implementation of dementia care robots are done in a way that has actual use for the people in the field and reduces the chance of unmet expectations, missed requirements and overpromises. Autonomy is an important factor in this since they, on average, relieved more of the burden on the caretakers. However, this is counteracted by their lesser acceptance by patients, who feel more comfortable if a family member or caretaker was in control. Autonomous robots hence did not decrease someone's level of loneliness to a similar level as a controlled robot does. Therefore, the level of autonomy is an important consideration while developing a social robot. 

• Daniel Woods, Fengpei Yuan, Ying-Ling Jao, and Xiaopeng Zhao. 2021. Social Robots for Older Adults with Dementia: A Narrative Review on Challenges & Future Directions. In Social Robotics: 13th International Conference, ICSR 2021, Singapore, Singapore,  November 10–13, 2021, Proceedings. Springer-Verlag, Berlin, Heidelberg, 411–420. https://doi.org/10.1007/978-3-030-90525-5_35

Social robot can improve social engagement, attention, neuropsychiatric symptions, reduce agitation, stress, increase neuron activity and decrease depression. One interface which could be applied are speaking agents such as Siri or Alexa. A Social Assistive Robot SAR is able to provide companionship, monitor health, simulate and provide therapy. The vast majority (> 80%) of the PwD and elderly are willing to take a SAR home and a similar percentage (> 75%) of caretakers would entrust someone to a SAR. Ethical concerns about reduced human interaction, loss of privacy, mismatched expectations and attachment to robots do exist however. Papers target companionship (41%), cognitive simulation (35%), monitoring (14%) and support for their impairment (10%). 60% of studied used small samples sizes and only ~15% of speech interfaces had a sample size larger than 40. All of those, however, where only run as one or two sessions. 40% of studies where done in a real-world setting (i.e. home or care centre) Only two large scale, real-world studies where conducted. Most paper where just proof-of-concepts where tests where run for a limited time, with a few sessions, low robot autonomy or in lab settings.

• M. R. Lima et al., "Conversational Affective Social Robots for Ageing and Dementia Support," in IEEE Transactions on Cognitive and Developmental Systems, vol. 14, no. 4, pp. 1378-1397, Dec. 2022, doi: 10.1109/TCDS.2021.3115228.

One of the earliest stages of dementia is Mild Cognitive Impairment which mildly impairs someone's memory, thinking and functioning. Cognitive games can be an effective strategy to mitigate the effect of MCI. This includes working memory, reasoning, attention, functioning, reduced anxiety and reduced depression. Robot interfaces are particularly potent interfaces for such games. This study focuses on getting feedback from the elderly and caretakers about cognitive games on robot interfaces. 10 participants perform a cognitive exercise on a robot after which they performed a questionnaire about the perceived enjoyment and difficulty of the game. Finally, this was supplemented by two experts giving their opinion of the game. Games were typically seen as easy to moderate with relatively high satisfaction, but feelings were mixed on whether they would like to do it again. Accent and word choice were also downsides. Finally, participants found the game boring and repetitive. Experts considered the concept of the game beneficial to people with MCI but saw problems in accessibility and relative complexity. Suggestions include making it easier for us by using selection rather than dragging. Including negative reinforcement caused the robot to sound patronizing and taunting.  Researchers noticed that participants struggled with understanding the instructions. The biggest issue with accessibility is touch screen sensitivity and finger pressure. 

• Mikaela Law, Ho Seok Ahn, Bruce MacDonald, Dina-Sara Vasjakovic, JongYoon Lim, Min Ho Lee, Craig Sutherland, Kathy Peri, Ngaire Kerse, and Elizabeth Broadbent. 2019. User Testing of Cognitive Training Games for People with Mild Cognitive Impairment: Design Implications. In Social Robotics: 11th International Conference, ICSR 2019, Madrid, Spain, November 26–29, 2019, Proceedings. Springer-Verlag, Berlin, Heidelberg, 464–473. https://doi.org/10.1007/978-3-030-35888-4_43

Application to our Robot
Prior work indicates the importance and potential of using social robots to engage patients with dementia, slow the process and improve the quality of their life. The highest need of the patient is that of getting individual care and combating the feeling of loneliness. Due to the progressive nature of dementia, it keeps asking for a greater role of the caretaker. Robots in this realm therefore should augment humans rather than replace them. Our robot, therefore, aims to ease the interactions between the closed one to the person in question to ease their ability to socialize with them. By allowing them to add their own information, it gives a personal connection to the robot while still maintaining the autonomy of the agent. This gives the biggest decrease in the workload of the caretaker. The choice of helping with communication is a good one since this is what the target population expects of the robot.