目的 针对我国老龄化背景下轻度认知障碍（MCI）老年群体面临的认知衰退与情感孤独双重挑战,现有认知干预产品多采用标准化游戏模式,缺乏个性化适应与自然交互特征。本研究探索大语言模型在老年认知陪伴领域的应用,设计基于智能对话的陪伴型机器人系统,为轻度认知障碍老年人提供个性化、情感化的认知刺激服务。方法 基于认知刺激理论与大语言模型能力的映射关系,通过对既有老年用户需求研究及文献的综合分析,提炼设计原则与功能指标;采用层次分析法（AHP）构建功能评估体系,计算权重值并完成一致性检验以确立设计优先级;通过产品定位、系统构架、功能模块与交互方案设计等流程完成设计实践。结果 本研究构建了名为“Echo”的认知刺激陪伴型机器人,采用多轮对话记忆机制和认知算法实现情境化交互与动态内容生成,通过随机性输出维持刺激内容的多样性与挑战性。系统以智能机器人为载体,提供自然语言形式的认知刺激服务,形成覆盖日常陪伴、非药物认知干预和情感支持的闭环模式。结论 基于大语言模型的陪伴型机器人突破了传统认知训练产品刺激方式单一、个性化不足及交互体验欠佳的局限,也实现了认知刺激与情感陪伴的深度融合,为轻度认知障碍老年人提供了更加自然、人性化的数字健康服务,对适老化智能产品设计具有参考价值。
Abstract
Against the backdrop of population aging in China, the elderly group with mild cognitive impairment (MCI) faces dual challenges of cognitive decline and social isolation and the existing cognitive intervention products mostly adopt standardized game-based models, lacking personalized adaptability and natural interaction features. Therefore, the work aims to explore the application of large language models in the field of cognitive companionship for the elderly, and design a conversational AI-powered companion robot system to provide personalized and emotion-oriented cognitive stimulation services for the elderly with MCI. Based on the mapping relationship between cognitive stimulation theory and the capabilities of large language models, design principles and functional indicators were extracted through a comprehensive analysis of existing research on elderly users' demands and relevant literature. The Analytic Hierarchy Process (AHP) was adopted to construct a functional evaluation system, calculate weight values, and complete consistency checks to determine design priorities. The design practice was finalized through a series of processes including product positioning, system architecture design, functional module design, and interaction scheme design. A cognitive stimulation companion robot named "Echo" was established, which adopted a multi-turn dialogue memory mechanism and cognitive algorithms to achieve contextualized interaction and dynamic content generation. Randomized output was incorporated to maintain the diversity and challenge of stimulation content. The companion robot based on large language models breaks through the limitations of traditional cognitive training products, such as a single mode of stimulation, insufficient personalization and suboptimal interaction experience. It also achieves the deep integration of cognitive stimulation and emotional companionship, thereby providing more natural and humanized digital health services for the elderly with MCI, which serves as a valuable reference for the design of aging-friendly intelligent products.
关键词
大语言模型（LLMs） /
陪伴机器人 /
认知刺激 /
轻度认知障碍（MCI） /
AHP
Key words
large language models (LLMs) /
companion robots /
cognitive stimulation /
mild cognitive impairment (MCI) /
AHP
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 国家统计局. 第七次全国人口普查公报(第五号): 人口年龄构成情况[EB/OL]. (2021-05-11)[2024-01-20]. http://www.stats.gov.cn/tjsj/tjgb/rkpcgb/qgrkpcgb/202106/t20210628_1818824.html
National Bureau of Statistics. The Seventh National Population Census Bulletin (No. 5): Population Age Composition[EB/OL]. (2021-05-11)[2024-01-20]. http://www.stats.gov.cn/tjsj/tjgb/rkpcgb/qgrkpcgb/202106/t20210628_1818824.html
[2] JIA L, DU Y, CHU L, et al.Prevalence, Risk Factors and Management of Dementia and Mild Cognitive Impairment in Adults Aged 60 Years or Older in China: a Cross-sectional Study[J]. The Lancet Neurology, 2020, 19(12): 987-995.
[3] OPENAI. GPT-4 Technical Report[R]. San Francisco: OpenAI, 2023.
[4] CLARE L, WOODS R T, MONIZ-COOK E D, et al. Cognitive Rehabilitation and Cognitive Training for Early-stage Alzheimer's Disease and Vascular Dementia[J]. Cochrane Database of Systematic Reviews, 2003(4): 32-40.
[5] LERMER E, STREICHER B, SAXL M, et al.Gamification in Cognitive Training: a Review on the State of the Art and the Role of Motivation[C]//Proceedings of the 8th International Conference on Pervasive Computing Technologies for Healthcare. Brussels: ICST, 2014: 399-402.
[6] PARK D C, BISCHOF G N.The Aging Mind: Neuroplasticity and Salutary Cognitive Engagement[J]. Annual Review of Psychology, 2013(64): 491-517.
[7] YI Z, OUYANG J, XU Z, et al. A Survey on Recent Advances in LLM-Based Multi-Turn Dialogue Systems[J/OL]. (2024-07-12)[2025-03-15]. https://arxiv.org/abs/2402.18013.
[8] YATES L A, YATES J, ORRELL M.Cognitive Stimulation Therapy for Dementia: History, Evolution and Internationalism[M]. London: Routledge, 2020.
[9] WOODS B, O'PHILBIN L, FARRELL E M, et al. Reminiscence Therapy for Dementia[J]. Cochrane Database of Systematic Reviews, 2018, 3(3): 112.
[10] FAZIO S, PACE D, FLINNER J, et al.The Fundamentals of Person-centered Care for Individuals with Dementia[J]. The Gerontologist, 2018, 58(Sup.1): 10-19.
[11] SPECTOR A, WOODS B, ORRELL M.Making a Difference: an Evidence-based Group Programme to Offer Cognitive Stimulation Therapy (CST) to People with Dementia[M]. London: Hawker Publications, 2006.
[12] WOODS B, AGUIRRE E, SPECTOR A, et al.Cognitive Stimulation to Improve Cognitive Functioning in People with Dementia[J]. Cochrane Database of Systematic Reviews, 2012(2): 55.
[13] NKETSIAH E, YEBOAH A, KUMI-YEBOAH A, et al.Incorporating Spirituality into Cognitive Stimulation Therapy Groups for Persons with Dementia: An Exploratory Study[J]. Journal of Gerontological Social Work, 2025, 68(1): 23-36.
[14] AGUIRRE E, SPECTOR A, HOARE Z, et al.Maintenance Cognitive Stimulation Therapy (CST) for Dementia: a Single-blind, Multi-centre, Randomized Controlled Trial of Maintenance CST vs. CST for Dementia[J]. Trials, 2010(11): 46.
[15] VAN RIJN H, VAN HOOF J, STAPPERS P J.Designing Leisure Products for People with Dementia: Developing "the Chitchatters" Game[J]. American Journal of Alzheimer's Disease & Other Dementias, 2010, 25(1): 74-89.
[16] ELLWARDT L, AARTSEN M, DEEG D, et al.Does Loneliness Mediate the Relation between Social Support and Cognitive Functioning in Later Life?[J]. Social Science & Medicine, 2013(98): 116-124.
[17] LIU Y, ZHANG S, TOMATA Y, et al. Emotional Support (Giving or Receiving) andRisk of Incident Dementia: the Ohsaki Cohort2006 Study[J]. Archives of Gerontology and Geriatrics, 2020(86): 103964.
[18] WEI J, TAY Y, BOMMASANI R, et al. Emergent Abilities of Large Language Models[J/OL]. Transactions on Machine Learning Research.(2022-04-19)[2025-04-29]. https://openreview.net/forum?id=yzkSU5zdwD.
[19] BOMMASANI R, HUDSON D A, ADELI E, et al. On the Opportunities and Risks of Foundation Models[R/OL]. Stanford: Stanford Institute for Human-Centered Artificial Intelligence.(2021-05-12)[2025-04-25]. https://arxiv.org/abs/2108.07258.
[20] JIANG J Y, WANG S, LI Q T, et al.A Cognitive Stimulation Dialogue System with Multi-Source Knowledge Fusion for Elders with Cognitive Impairment[C]//Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics. Toronto: Association for Computational Linguistics,2023: 10628-10640.
[21] 白惠天, 王强, 吴朋阳, 等. 十问“AI陪伴”:现状、趋势与机会[EB/OL]. (2024-10-25)[2025-06-28]. https://tisi.org/30696/.
BAI H T, WANG Q, WU P Y, et al. Ten Questions about "AI Companionship": Status, Trends,Opportunities[EB/OL]. (2024-10-25)[2025-06-28]. https://tisi.org/30696/.
[22] NADERZADEH M, ARABALIBEIK H, MONAZZAM M R, et al.Comparative Analysis of AHP-TOPSIS and Fuzzy AHP Models in Selecting Appropriate Nanocomposites for Environmental Noise Barrier Applications[J]. Fluctuation and Noise Letters, 2017, 16(4): 1750038.
[23] 杜栋, 庞庆华. 现代综合评价方法与案例精选[M]. 第3版. 北京: 清华大学出版社, 2016: 67-75.
DU D, PANG Q H.Modern Comprehensive Evaluation Methods and Selected Cases[M]. 3rd, ed. Beijing: Tsinghua University Press, 2016: 67-75.
[24] SAATY T L.Decision Making with the Analytic Hierarchy Process[J]. International Journal of Services Sciences, 2008, 1(1): 83-98.
基金
四川美术学院博士科研启动项目（22BSQD017）; 科普资源的数字化与产业化建设（HZ2021010）
PDF(668 KB)
渝公网安备 50010702501717号