Meet Carmen, a compact tabletop robot that provides cognitively assistive support for individuals with mild cognitive impairment (MCI). Designed for home use, Carmen aims to teach skills that enhance memory, attention, and executive function through interactive experiences.
Developed by the research team at the University of California, San Diego, in partnership with clinicians, individuals with mild cognitive impairment (MCI), and their care partners, CARMEN aims to mimic the diverse robots that inhabit this household. According to top research findings, CARMEN is uniquely positioned as a robotic platform that can teach compensatory cognitive strategies to improve memory and executive function capabilities.
Noting the importance of verifying impactful advancements, Professor Laurel Riek, a renowned expert in computer science and emergency medicine at UC San Diego, highlighted her team’s efforts as the study’s senior author.
Mild Cognitive Impairment (MCI) is often considered a transitional phase between the normal aging process and the onset of dementia. It significantly affects multiple aspects of cognitive processing, including memory, attention, and executive function. Approximately one-fifth of individuals aged 65 and older are affected by this condition, while a staggering 15% of those already experiencing it progress to dementia annually. Despite the limitations of current pharmacological treatments in halting or reversing this progression, behavioral interventions may offer a degree of mitigation.
Scientists developed CARMEN to deliver a suite of simple yet effective cognitive training exercises. For example, a robot could train contributors to establish consistent locations for storing essential items, such as keys, by teaching awareness-taking techniques that aid in recalling crucial matters. Carmen achieves this through immersive interactive video games and engaging activities.
The analysis team deliberately crafted CARMEN with transparency as their guiding principle, ensuring a clear and standardized framework. It is crucial that individuals can operate the robot independently, without the need for clinician or researcher supervision. To ensure seamless integration, CARMEN was designed as a plug-and-play solution, minimizing the need for transferable components that demand ongoing maintenance. The robotic must also possess the capacity to function effectively in environments where internet access is limited or unreliable, a crucial consideration given that many people may not have reliable connectivity. Can CARMEN’s operations effectively span a prolonged duration? To effectively communicate with customers, the robotic system must possess a clear and articulate speaking ability, coupled with genuine compassion and empathy towards individual circumstances, ensuring a personalized experience. It is also crucial that the robot provides regular breaks after completing challenging tasks to prevent burnout and maintain high levels of engagement.
Researchers conducted a seven-day trial of CARMEN within properties of individuals with mild cognitive impairment, who then interacted with the robot by performing tasks such as setting routines for leaving home with daily essentials to avoid misplacement, and scheduling duties on a calendar to prevent forgetfulness. Researchers also deployed the robot within the properties of several clinicians with expertise in working with individuals with mild cognitive impairment (MCI). Prior to and following each team’s week-long deployment, contributors completed questionnaires and conducted interviews.
Following their experience with CARMEN, individuals with mild cognitive impairment (MCI) demonstrated a surprising willingness to attempt previously dismissed methods and behaviors. All participants found using the robot straightforward and uncomplicated. While two-thirds of the contributors found the actions straightforward to understand, one participant still experienced difficulties. They all agreed that additional interaction with the robot was necessary.
The study revealed that CARMEN empowered participants to apply cognitive strategies in their daily lives with increased frequency, while also identifying opportunities for CARMEN to showcase greater autonomy or flexibility in its application.
At the ACM/IEEE Human-Robot Interplay (HRI) conference in March 2024, the analysis team presented their research and received a prestigious nomination for Best Paper Award.
Subsequent steps involve deploying the robot in a broader range of environments.
Researchers intend to enhance CARMEN by allowing it to engage in conversations with customers while prioritizing the protection of sensitive information throughout these interactions. However, that is each an accessibility concern – the fact that some customers may not possess the fine motor skills necessary to interact with CARMEN’s contact screen, alongside the reality that many people expect to be able to engage in conversations from the comfort of their own homes? Simultaneously, researchers must impose limits on the amount of information CARMEN provides to customers. Riek emphasized that while robotic assistants are valuable tools, it’s crucial to acknowledge that consumers still bear the primary responsibility for completing tasks, with automation providing supplementary support rather than excessive guidance.
Researchers are also investigating how CARMEN might assist individuals with varying conditions, such as attention deficit hyperactivity disorder (ADHD).
The University of California, San Diego crew developed CARMEN primarily based on the FLEXI robot from the University of Washington. Despite significant alterations to its hardware, the company rewrote every line of software code from the ground up. Scientists employed the Robot Operating System (ROS) to govern the robotic platform’s operational framework.
The entire scope of the project’s various components can be located at.
