Toward Culture-Aware Elderly Care Robots. Nak Young Chong School of Information Science Japan Advanced Institute of Science and Technology

Transcription

1 Toward Culture-Aware Elderly Care Robots Nak Young Chong School of Information Science Japan Advanced Institute of Science and Technology

2 One of the most homogeneous ethnicities in the world Changing demographics Preparing for the challenges of a nonhomogeneous society Japan: Super-aging Society

3 Choose the right action Choose the right topic of conversation Know the environment, furniture and objects. Other Examples of practical use of cultural competence

4 Transcultural Robotic Nursing Cultural Knowledge Representation Culturally-sensitive Planning and Execution Culture-Aware Interaction in a Smart ICT Environment

5 Cultural Knowledge Base Long term information about the user Cultural knowledge Culturally-sensitive Planning & Execution Module CAHRIM Current state & User requests Actions to be executed Culture-Aware Human-Robot Interaction Module Executor

6 Navigation Emotion Generation Emotional Body Expression Human Emotion Recognition Verbal and touch-based interaction Monitoring and control of ihouse functions Human activity recognition Basic motion/perceptual capability expansion

7 EU Consortium Cultural background information uaal-echonet interface ihouse (ECHONET) CARESSES Robot Individual information Personality traits Culturally competent services Culture-Aware Verbal/ Nonverbal Human-Robot Interaction Smart ICT Environment User ADL data logs Culturally competent services Human emotion/intention estimation

8 obstacle avoidance, safe trajectory, finding free space, and exploration and localization. Navigation

9 Navigation: SLAM

10 The interrelated observations and theories of man s use of space as a specialized elaboration of culture. [Edward T. Hall] Navigation: Proxemics

11 Navigation: Proxemics

12 Long-term Memory Environment Sensory Representation Object-based Visual Features Semantic Memory Storage Filter Scene-based Visual Features Episodic Memory Human Social Referencing/ Sharing Emotion Model Emotion generation Memory Retrieval w/ Forgetting Mechanism Time Decay Memory Consolidation Personal Emotion

13 α Arousal (Fast) γ γ (Negative) β Valence (Positive) α (Slow) Emotional body expression 13

14 Emotional body expression

15 Phase 1: Learning to control emotions aligned with a culture Phase 2: Controlling emotions adapting to a new culture Emotional body expression 15

16 Emotional body expression 16

17 Descriptor Descriptor Descriptor representative Descriptor Descriptor set of actions descriptor representation Emotional body expression clustering imitation

18 Emotional body expression

19 Audio Facial expression Wearable sensors Human emotion recognition

20 Ms. Yamada pre lunch routine

21 ihouse ECHONEToriented Objects Actuator Device Sensor Appliance CARESSES Robot Robot Controller uaal Middleware SB CB CARESSES Cloud uaal Middleware SB CB echowand ECHONET API Home Gateway uaal Middleware SB CB Controller uaal Adaptation RDF Objects ECHONET Objects Database Syntax Translation uaal Objects uaal Ontologies Data Transferring Context Bus (CB) Service Bus (SB)

22 ihouse NLP library Robot Application Data Transferring Context Bus Service Bus ECHONET based Devices Actuator Device Sensor Appliance Home Gateway User NAOqi Framework Cloud Server universaal Middleware ECHONET Interface echowand Robot Controller universaal Middleware universaal Middleware Controller RDF Objects uaal Objects uaal Adaptation Syntax Translation ECHONET Objects uaal Ontologies Database ihouse real time monitoring and control

23 ihouse real time monitoring and control

24 keywordextraction Verbal and touch-based interaction

25 Activities Recognition rates for Activities [%] Sleeping Studying Cooking Eating Watching TV Activity recognition

26 Advinia Health Care Homes, UK Clients from 2 cultural groups (white English and Indians) 10 Clients per cultural group 2 weeks per client Hisuisui, Ishikawa, Japan Clients from 1 cultural group (Japanese) 10 Clients per cultural group 2 weeks per client Testing and Assessment

27 Cultural identity components and assessment Heterogeneous ICT network-care robot integration Human-care robot interaction evaluation: What are good/ bad practices? Concluding remarks