Establishing contextually aware sensors for elderly activity monitoring systems

Homecare systems have become a focus of research due to shifting care requirements of the elderly, with the home becoming a centre of care provision. Hospitals are reaching their capacity to provide services to this growing elderly population, and with the number of homecare professionals dwindling, solutions are required. Changing requirements, due to the increasing prevalence of non-communicable diseases, leads research to examine lifestyle metrics such as activity while monitoring home patients. Devices used to monitor these activities described in the literature however, express issues in establishing consistent activity determination, distinguishing between the contextually ambiguous actions of an occupant and ambiguity on data privacy. This research describes an information health system (IHS) to address the potential gaps in the area which are in line with shifting care requirements and a greater focus on the informal care provider in the absence of professional providers. The proposed care system will seek input from these stakeholders using surveys in a bid to establish overall feasibility in the space. Monitoring is achieved by creating a model of patient care requirements. Experimentation will examine monitoring devices as a capable standard to not only locate occupants within a home with a verifiable degree of accuracy but to determine care outcomes by explaining context within these locations using ADL and the results from the key stakeholder survey. Data captured is evaluated to determine activity and form a sensor driven care model, captured data is stored locally to alleviate privacy concerns associated with sensor device limitations. Contextual awareness within this data is examined by observing occupants interacting with ‘unique’ objects of interest (OOI) or the change of on-body sensor ques, such as posture or heart rate. Sensors are shown to be effaceable in this environment while fulfilling the stated care requirements of ADL, additional experiments seek to establish privacy. Novel validation methods were used to classify activity captured using the examined IHS devices. From these experiments, this research will attempt to establish sensor driven devices as the optimal approach to homecare due to the accuracy, novel examination of context, interoperability and low maintenance.