From speech acts to assistance acts for cognitive assistance in ambient assisted living: how to nudge cognitively impaired people to act independently

Technology-driven cognitive assistance is often either excessive or insufficient, and then not necessarily helpful, because of the inadaptability to the needs and abilities of people with cognitive impairments. Cognitive assistance is usually a dialogue (an interaction) where the caregiver provides assistive cues to which the assisted person replies with behaviours. This research proposes an ontology-based cognitive assistance model for ambient assisted living (AAL) systems. It aims to characterize the situations where assistance is needed and to provide an adaptive dialogue between the AAL system and individuals to help them cope with problematic situations in a minimally assisted manner, in so far as that is possible. The theory of speech acts is extended to translate cognitive assistance into adaptive assistive messages. The model enables AAL systems to identify when to assist according to the sensor observations and to deliver progressive minimal guidance messages with stops, or adjustments and replays, depending on the behaviours of people living with cognitive impairments, such as traumatic brain injury. Such user-centered assistance considers both the abilities, and the assistance needs of individuals, to give the right level of assistance that just nudges them towards carrying out tasks “on their own”. The model builds on cognitive assistance as used within clinical practice in the field of TBI and extends it to enable AAL systems to deliver assistance acts in due course through various actuators during the person’s everyday life activities. The model is formalized in an ontology to enable reasoning capability, extension, and reuse in different AAL systems. The illustration using a real-life scenario shows its usability and applicability in the provision of cognitive assistance through graded guidance message spreading.