Dem@Care Research Activities

This activity focuses on the definition of user-level requirements and the evaluation of the impact of Dem@Care intervention on people with dementia, their carers and clinicians responsible for their treatment. The Medical team of the project will determine the clinical requirements and sensor measurements necessary to meet the needs of the people with dementia and their caregivers. This will define the scenarios to be implemented by the system, involving situations that will provide meaningful information about the person’s condition, without interfering with their daily life. The safeguarding of ethics and confidentiality will also be defined, according to regulations and legislation in each country. Finally, the impact of the system on societal and personal levels will be analysed.

Within the framework of Dem@Care, physiological and lifestyle information will be collected from a multitude of sensors. Unobtrusive sensors for physiological measurements will be deployed to provide a constant and accurate picture of the health status of the person with dementia, to be communicated to their care provider. Lifestyle sensors will be used to monitor lifestyle patterns like electricity consumption which tend to be unchanging when the person’s life is normal. Early fusion will allow the extraction of conclusions concerning the person’s state by mining the physiological and lifestyle data that can be used for the early detection of unusual events or patterns.

This activity of Dem@Care focuses on the analysis of daily activities of the people with dementia in their domestic environment via the extraction and processing of features describing their actions and the context in which they occur. Data -driven descriptors will be extracted from various raw contextual and wearable sensors and their interpretation in terms of activities and events. This work complements the extraction of health and lifestyle features, and provides faceted features used in Medical Ambient Intelligence. The set of complementary features consists of visual perception for the detection and recognition of situations of interest, audio-based affective analysis for characterizing the person’s behavioural, mental and emotional state, instrumental activities monitoring for characterizing the person’s actions and the organization of detected events into a life-log suitable for further analysis.

The objective is to provide the knowledge structures and intelligent analysis and decision making procedures that will enable the customised interpretation of a person’s behaviour and the delivery of appropriately tailored management and support services, in line with the purposes of the use cases addressed within Dem@Care.

To this end, this project activity addresses:

• The representation and linking of clinical and generic knowledge pertinent to the semantic fusion and high-level interpretation of the faceted descriptions extracted through the multi-sensor infrastructure.
• The extraction and maintenance of time-evolving behavioural patterns.
• The customised update and enrichment of background knowledge in a consistency preserving manner.
• The development of practical reasoning techniques for the integrated interpretation of a person’s behaviour and for the inference of appropriate feedback.

The development of techniques that can scale to the volumes of data considered will be investigated, while effectively coping with uncertainty, and with incomplete and noisy input will be also explored.

Information and necessary feedback needs to be provided to people with dementia and clinicians in order to realise an integrated solution for the remote management and treatment of dementia. Client-side Interaction research area investigates how best to provide system interaction for the respective groups in order to achieve the desired result. This includes human-computer-interaction as well as human-to-human interaction such as between the clinician and the person with dementia. The main axes are:

• Provide the necessary feedback on time to persons during their daily life and give reminders, warnings and information only when suitable and adapted to situation.
• Ensure interaction with the Intelligent Home environment during daily routines for people with dementia.
• Provide routines and protocols that will allow the interaction of clinicians with people with dementia.

The study in this area will result in the development of the architecture and the integration of Dem@Care system by unifying the concepts, algorithms and techniques developed in the other WPs with the sensors and components.

More specifically, the objectives are:

• To translate the defined requirements and scenarios into real functionalities, to be implemented in the prototypes.
• To define the system architecture, communication data and interfaces by taking into account the developed algorithms and techniques and integrate each prototype according to the feasibility study results.
• To perform data collection tests and system tests.
• To define the data privacy and security levels needed by the system and apply these results to the system architecture.

Validation and assessment the Dem@Care technology is an important aspect of the project. A three-staged evaluation process has been defined to ensure the saliency and effectiveness of the developed systems.

More specifically:

• For the first Dem@Care prototype: the first evaluation phase will verify usability, functionality and reliability, and will further refine the functional requirements.
• For the second Dem@Care prototype: the second evaluation phase will focus mainly on Dem@Care’s external qualities, and to the formative evaluation of suitability, accuracy, security, maturity, etc, among others.
• For the final Dem@Care prototype: the third and final evaluation phase will evaluate the overall efficacy and impact of the Dem@Care system, including clinical considerations (cognitive, behavioural, etc.), factor related to enhanced quality of life for people with dementia and their caregivers, and so forth.