Pervasive Health Workshop on Affective Interaction with Virtual Assistants within the Healthcare Context'

Research Article

Analyzing Drivers’ Affect for the Design of Intelligent Wheelchairs for Older Adults with Cognitive Impairment

  • @INPROCEEDINGS{10.4108/eai.16-5-2016.2263869,
        author={Genevi\'{e}ve Foley and Pooja Viswanathan and Ellen Zambalde and Alex Mihailidis},
        title={Analyzing Drivers’ Affect for the Design of Intelligent Wheelchairs for Older Adults with Cognitive Impairment},
        proceedings={Pervasive Health Workshop on Affective Interaction with Virtual Assistants within the Healthcare Context'},
        publisher={ACM},
        proceedings_a={AIVAHC2016},
        year={2016},
        month={6},
        keywords={intelligent wheelchairs human-robot interaction affect discomfort conflict},
        doi={10.4108/eai.16-5-2016.2263869}
    }
    
  • Geneviève Foley
    Pooja Viswanathan
    Ellen Zambalde
    Alex Mihailidis
    Year: 2016
    Analyzing Drivers’ Affect for the Design of Intelligent Wheelchairs for Older Adults with Cognitive Impairment
    AIVAHC2016
    EAI
    DOI: 10.4108/eai.16-5-2016.2263869
Geneviève Foley1,*, Pooja Viswanathan1, Ellen Zambalde1, Alex Mihailidis1
  • 1: University of Toronto
*Contact email: genevieve.foley@mail.utoronto.ca

Abstract

Potential powered wheelchair users may be denied access out of safety concerns due to cognitive, physical or perceptual impairments. To restore independent mobility, wheelchairs that provide driving assistance are under development. To establish requirements for these smart wheelchairs, 10 older adults with mobility and cognitive impairments were asked to drive in five scenarios, with a mock intelligent wheelchair operated by a tele-operator providing three different types of assistance, ranging from simple collision avoidance to fully autonomous driving. Videos of the trials were coded to extract participants’ affect during the tasks in relation with usability issues identified during the analysis of the interviews and questionnaire results. Affect with regards to two key usability issues was considered: (1) discomfort as an indicator of safety concerns and (2) conflict as an indicator of mismatched user expectations and system behavior. Instances of conflict were inversely associated with the level of tele-operator intervention. Discomfort levels were generally low with the exception of one scenario/control mode combination where fear of injury was noted.