tno human factors/ delft university of technology distributed embodied epartners in a ubiquitous...
TRANSCRIPT
TNO Human Factors/ Delft University of Technology
Distributed embodied ePartners in a ubiquitous computing environment
Mark Neerincx et al.
Human Factors
Introduction
Environments with more networked information
compilations and technical equipment.
Add a collection of distributed & connected
personal electronic partners, ePartners, to support
(distributed) human actors for specific tasks, like• health-care actions by diabetics, • technology use by elderly, and • disclosure of feelings in high-demand missions
Human Factors
ePartner: Human-Machine Collaboration
• Predicting the actors’ momentary needs• on-line gathering and modeling of human, machine, task and
context information
• Attuning the interaction to these needs• (semi-)automatic tailoring of support, content and dialogue to
establish optimal human-machine performance
• Establishing “natural” H-M communication• expressing and interpreting communicative acts based on a
common reference
Human Factors
Personal ePartner
• Sharing of experiences, learning and anticipation• Common goal, but different roles and responsibilities• Adequate distribution of workload • Not copy of a human, but a complement• Taking care of individual characteristics and context of operation• Adequate level of trust and affection• Quick understanding and natural communication (“one word is
enough”)
for specific goal and domain (e.g. self-care, daily
activities)
Human Factors
iCat as ePartner
TNO and DUT use the iCat to develop • models & prototypes for effective & social H-M
communication.
Research questions centre on • the sensing and generation of affective expressions
(face, voice), • the application of different communication and
assistance styles like “motivational interviewing” and
“cooperative anamnesis”, • the effects of the embodiment of an ePartner (e.g.
compared to a virtual character).
Human Factors
The SuperAssist Project
Personal assistants supporting distributed supervision ofcomplex task environments. Each human actor has an assistant for anomaly detection, diagnosis and communication
Human Factors
Focus on Tele- and Self-Care
• More patient’s involvement in the care process • Operating domestic medical instruments • Management of a computer-based patient record• Communication with medical & technical specialists
• Increasing number of older people• Chronic diseases: 50% non-adherence
• Diabetes Type II:• Large, increasing group• Typically older than 40• Pancreas produces too little insulin• Treatment involves diet, no-smoking, medication &
exercising
Human Factors
Bottlenecks in Diabetes Care
• Shortage of diabetes nurses
• High No-Show in diabetes policlinic
• Lack of information during consultation
• Fixed consultation frequency (every 3 months)
• Lack of involvement/motivation of patients
• Too brief consult for adequate feedback to patients
• Large increase in patient numbers
Human Factors
Supporting the Patient
• Usage of medical equipment• Human-assistant cooperation
• Health and lifestyle• Therapy adherence
• Automation of assistance• Human vs synthetic assistant
• Embodiment of the Assistant• Social behavior
• Tests should show improvements for• Patient behavior and opinion (e.g. trust)• Adequacy and “richness” of dialogue
Human Factors
First iCat Experiment
Goal: • Guidelines for adherence support• iCat alternative for “desk-top” interface
Background:• Health Buddy®
(Health Hero Network)
Human Factors
Guidelines
• Personal assistants• Fun• Trust• Cooperation
• Psychology/ motivational interviewing• Express empathy• Communicate respect• Listen rather than tell
• Social robots• Express and perceive emotions• Communicate with high-level dialogue• Use natural cues (gaze, gestures, etc.)
Human Factors
Three User Interfaces
Tiggie (Doellgroup) iCatText (“chat”)
Human Factors
Pilot Test
Social & non-social iCat or Tiggie, vs text-interface:• 6 participants• Scenarios:
• 3 types (diet, self-check, medication), but similar• Both open and closed questions• 4 question blocks: Monday, Wednesday, Friday and Sunday
First results: Social iCat preferred• Trust• Acceptance• Empathy• Personality
Human Factors
Conclusions
Guidelines for Health Assistance• Trust • Fun• Cooperation• Express Empathy• Communicate respect• Listen rather than tell
Future experiments: • Elderly• Virtual iCat• Personalization• Dialogue management• Long-term effects• multi-Human – multi-ePartner
collaboration
Human Factors
Acknowledgement
• SuperAssist is partially funded by IOP-MMI Senter a program of the Dutch Ministry of Economics
• Three research institutes participate in the SuperAssist project: TNO, Delft University of Technology (DUT), and Leiden University Medical Center (LUMC)
• Four companies are involved: Philips Research, Sigmax, Pemstar, Science & Technology
Specific contributions to this presentation:• Rosemarijn Looije (RUG/TNO)• Amy Ahluwalia (LUMC)• Olivier Blanson Henkemans (DUT/TNO)• Jasper Lindenberg (TNO)• Lennard Kuijten (UM/TNO)• Vanessa Sawirjo (DUT)
Human Factors
Questions?