IEEE Robotics & Automation Magazine - March 2013 - 75

coaching for cognitive interventions is already lacking and on
a recognized steady decline [5]. There exists an urgent need to
further investigate the potential use of cognitive training interventions as a tool to aid the elderly.
The goal of our research is to advance knowledge in cognitive/social interventions for elderly individuals suffering from
cognitive impairments via the development of robotic technology [6], [7]. We aim to design humanlike, socially assistive
robots capable of providing cognitive assistance and social
interaction in self-maintenance (i.e., eating, dressing, and
grooming) and activities of daily living (i.e., cognitively and
socially stimulating leisure activities). These robots focus on
the core impairments of dementia and the ability to support
working memory, attention, awareness, and focus on task
behavior, to reduce a person's dependence on caregivers and
provide him/her social interaction during the course of these
activities. Our long-term goal is to study how such robots can
contribute to therapeutic protocols aimed at improving or
maintaining residual social, cognitive, and global functioning
in persons suffering from dementia.
In this article, we present the development of our unique,
expressive, humanlike, socially assistive robot, Brian 2.1
(Figure 1). Brian 2.1 can engage elderly individuals in both
self-maintenance and cognitively stimulating leisure activities.
The robot is able to determine its appropriate assistive behaviors based on the state of the activity and a person's user state.
The social abilities of the robot play an important role in creating engaging and motivating interactions customized for
individual users. In addition, we present the results of
human-robot interaction (HRI) studies conducted with

Two Nintendo
Wii Remote
IR Cameras

10-MP
Creative
Webcam

Kinect
Sensor

Figure 1. The expressive, humanlike, socially assistive robot,
Brian 2.1.

elderly users at a long-term care facility to investigate the
overall acceptability of such a robot for the intended activities.
Namely, the study consisted of observing user engagement
and compliance during interactions with Brian 2.1 as well as
obtaining user feedback regarding acceptance of the robot as
an assistive tool via a questionnaire administered after the
interactions.
Related Work
To date, limited research has been conducted on the use and
benefits of social robots as therapeutic aids or assistants for
the elderly. However,
such robots have the
potential to bring a new
The number of older adults
interaction tool to a vulnerable population that
who have difficulties
would otherwise lack
resources. They also
performing self-care
appear to make existing
care cheaper and more
and independent-living
effective [8]. For example, the seal-like robot
activities increases
Paro has been designed
to engage the elderly in
significantly with the
animal therapy by learning which behaviors are
prevalence of cognitive
desired from the way a
person pets, holds, or
impairment.
speaks to it. Studies performed with Paro have
shown that the robot can improve users' mood and stress
levels as well as facilitate the interaction between users by
creating a comfortable and sociable atmosphere [9]. The
Pearl robot was developed to perform tasks in assisted living
facilities, such as providing a person with reminders, guiding him/her to appointments, and providing information
assistance [10]. Experiments with a group of six elderly people verified the robot's ability to autonomously and effectively complete the guidance task. A music game study
performed with the childlike robot Bandit II and three
elderly participants with cognitive impairments showed an
improvement in cognitive attention and task performance
over a six-month period [11]. The same robot was also used
in [12] as an exercise instructor. A feasibility study with the
11 elderly participants showed that the robot could motivate
these individuals to perform simple physical exercises.
The novelty of Brian 2.1 with respect to the aforementioned robots is its increased humanlike social abilities.
Namely, the robot can determine both user engagement and
activity state during HRI and, in turn, use this information in
real time to determine its own emotional assistive behaviors.
We also propose two new assistive interaction activities for
the robot: eating and playing a card game.
To date, the majority of outcomes that have been utilized
to study HRI with interactive robots and the elderly have
focused primarily on task performance. A handful of studies
March 2013

IEEE rOBOTIcS & aUTOMaTION MaGaZINE

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - March 2013