SHB: Small: Socially Assistive Human-Machine Interaction for Improved Compliance and Health Outcomes

世界人口正在老龄化；据估计，到2050年，85岁以上的人口将是现在的三倍。许多人预计需要身体和认知方面的帮助，所有人都需要额外的医疗保健。初级保健医生、护士和管理护理设施空间和人员的短缺已经成为当今的一个问题，这为辅助技术填补护理空白创造了一个利基市场。越来越多的研究表明，某些行为模式对长寿和健康有积极影响，包括定期体育锻炼、社交互动和认知参与。该项目旨在开发和评估社会辅助人机交互技术，以影响用户参与和遵守家庭或养老院的健康促进行为，以延长寿命和提高生活质量。这项工作的重点是开发社会辅助系统（SAS），使用对等水平的人机交互，其中机器充当知识丰富的具身代理的角色，能够提供长时间的、持续的和引人入胜的交互。这项研究的重点是两种促进健康的人机交互：1)锻炼课程（认知和/或身体）和2)社交课程。对于一对一的锻炼课程，该项目正在开发SAS功能，提供锻炼监测、指导和激励。在社交方面，该研究正在开发SAS功能，提供社交互动和友好的提醒和鼓励，以遵守健康养生法（吃药、锻炼身体）和健康习惯（给家人打电话、见朋友）。这两种交互具有影响人类行为的共同目标，研究的主要贡献是通过人机交互影响行为的一套方法和算法。这是通过一种新的双重方法实现的：1)短期互动的自适应互动指导，以及2)长期互动中的行为激励和指导，以保持参与并提高任务绩效。SAS的体现是为了最大限度地提高参与度和合规性；这项工作的一个重点是开发一种自然、有说服力和引人入胜的具体沟通的综合方法。正在开发的方法和算法是通用的，将在基于计算机和基于机器人的代理上实施和测试。通过这种方式，本工作在多个技术平台和实施例上验证所开发的方法，并比较其相对有效性和预期用户群体的可接受性。实现的细节是由早期的焦点小组告知老年人参与者，目标用户群体。开发的技术正在人机交互（HCI）和人机交互（HRI）版本中实施，并正在通过一项为期数周的用户研究，与SAS实现交互的一大组老年用户进行评估。这项工作的重点是生成方法、算法和大量多模态数据（视频、音频和问卷），为健康和健康相关的HRI和HCI的持续研究奠定基础。该项目旨在解决全国公认的促进老龄化人口健康和长寿的医疗保健挑战的一个组成部分。除了基本的算法和方法开发之外，该项目还实现和测试了现实世界的社会辅助系统，包括基于计算机和基于机器人的系统，其中包括大量老年养老院居民。该项目有望为研究和医疗保健产品开发提供有用的见解，以解决这一不断增长的人口部分。除了社会相关的研究重点外，项目团队还参与了一项全面的K-12外展活动计划，该计划利用机器人技术促进STEM主题学习，并利用提案的健康主题。该项目涉及市中心K-12的教师和学生，每年开放房屋，集会和研讨会，为老年人提供辅助系统的实践经验，以及继续学习的材料。项目团队包括博士生、本科生和K-12志愿者，并建立了一个指导管道，使大学生既能得到指导，又能成为年轻同龄人的导师和榜样。

The world's population is growing older; it is estimated that in 2050 there will be three times more people over the age 85 than there are today. Many are expected to need physical and cognitive assistance, and all will need additional healthcare. Shortages in primary care physicians, nurses, and managed care facility space and staff are already an issue today, creating a niche for assistive technologies to fill the care gap. A growing body of research shows that certain behavior patterns have a positive impact on longevity and wellness, including regular physical exercise, social interaction, and cognitive engagement. This project aims to develop and evaluate socially assistive human-machine interaction techniques that influence the user to engage in and comply with wellness-promoting behaviors in the home or nursing home in order to enhance longevity and quality of life.This work is focused on developing socially assistive systems (SAS) using peer-level human-machine interaction in which the machine serves in the role of a knowledgeable embodied agent capable of providing time-extended, sustained, and engaging interaction. The research focuses on two types of wellness-promoting human-machine interactions: 1) exercise sessions (cognitive and/or physical) and 2) socializing sessions. For one-on-one exercise sessions, the project is developing SAS capabilities that provide exercise monitoring, coaching, and motivation. For socialization, the research is developing SAS capabilities that provide social interaction and friendly reminders and encouragement to comply with health regimens (taking medicine, being physically active) and healthy habits (calling family, meeting friends). The two types of interactions share the common goal of influencing human behavior, and the main contribution of the research is the set of methods and algorithms for influencing behavior though human-machine interaction. This is achieved through a novel two-fold approach: 1) adaptive interaction steering for short-term interactions, and 2) motivation and coaching of behavior in longer-term interactions to maintain engagement and enhance task performance. The embodiment of the SAS is leveraged to maximize engagement and compliance; a key focus of the work is on developing a comprehensive method for natural, persuasive and engaging embodied communication. The methods and algorithms being developed are general, and will be implemented and tested on both computer-based and robot-based agents. In this way the work validates the developed methods on multiple technology platforms and embodiments, and compares their relative effectiveness and acceptability by the intended user population. The specifics of the implementation are informed by an early focus group with elderly participants, the target user population. The developed techniques are being implemented in both human-computer interaction (HCI) and human-robot interaction (HRI) versions, and are being evaluated with a large group of elderly users interacting with the SAS implementations over a multi-week user study. The work is focused on generating methods, algorithms, and a large corpus of multi-modal data (video, audio, and questionnaires) for grounding continued research into health and wellness-relevant HRI and HCI. This project aims to address a component of the nationally recognized healthcare challenge of promoting wellness and longevity in the aging population. Beyond basic algorithm and method development, the project implements and tests real-world socially assistive systems, both computer-based and robot-based, with a large population of elderly retirement home residents. The project is expected to produce insights useful for both research and healthcare product development for addressing this growing segment of the population. In addition to the socially relevant research focus, the project team is also engaged in a comprehensive program of K-12 outreach activities, which use robotics to promote STEM topic learning using the health theme of the proposal. The project involves inner city K-12 teachers and students in annual open houses, assemblies, and workshops that provide hands-on experiences with assistive systems for the elderly, as well as take-home materials for continued learning. The project team includes PhD students, undergraduates, and K-12 volunteers, and establishes a mentoring pipeline so that university students are both mentored and serve as mentors and role models for their younger peers.