Skip to main content

Supporting Active Ageing through Multimodal coaching

Periodic Reporting for period 2 - SAAM (Supporting Active Ageing through Multimodal coaching)

Reporting period: 2019-04-01 to 2021-05-31

As in the rest of the world, Europe faces an on-going crisis in caring for its aging population. Citizens are living much longer than ever before, with increasingly complex medical, social, and infrastructural needs, but most EU social support services and structures are lagging very much behind the growing need. While many efforts have focused on better access to healthcare and the expansion of assisted and social care housing, one of the main intervention points, enabling the aging population to remain in their homes longer, is just now becoming a primary developmental priority. Caring for the aging population as they transition from active and independent lifestyles to those requiring significant external support is a large and complex field, with many approaches to the great diversity of citizen need profiles.

SAAM is a large and ambitious project, focused on a single goal: providing new methods supporting Europe’s various ageing population to remain active and live independently for as long as possible. For this reason, all of the project’s objectives are grounded in practical user needs, and focused on the innovative methods most likely to find widespread use for maximum impact.

The project also focuses on addressing the needs of a particular segment of the population, the elderly above the age of 60 in economically under-developed regions with low technological proficiencies, who are most vulnerable in terms of access to support services, economic deprivation and social isolation.
The first important phase of the project was to investigate user needs through the cohort of the project’s social partners – SOCA, BRC, Caritas, EURAG. The four partners represent different support networks for the elderly – from formal and informal social (BRC, CAR) and health service providers (SOCA), to representatives of senior groups (EURAG, BRC). The investigation consisted of both online/phone and face-to face interviews with elderly representatives, followed by focus groups and informative interviews from stakeholders. A total of 506 questionnaires were gathered and analysed. As a result of the initial study basic user profiles and conceptual scenarios for the system functioning were elaborated as deliverables.

High Ethical standards were considered from the beginning of the project. Internal ethical boards were set-up for the project. All research, involving user interaction and study so far was subject to ethical approval both from the partner’s internal ethical committees and by the National Ethical Committee in Slovenia.

The technical requirements and architecture were specified, following the results of the initial study on user needs and requirements. Work has been organized in eight assessment and coaching domains, four being part of the basic system (Activity, Mobility, Sleep and Social Activity) and four as advanced system (Cardio-vascular health, Cognitive decline, Nutrition, and Language based emotional state).

In the first half of the project a hardware architecture was developed that enables collection and manipulation of required data, as well as a concept of coaching models (pipelines). Each coaching pipeline will take care of a specific type of coaching in a specific domain, such as coaching for proper standing up from a chair in the mobility domain. The first component of a coaching pipeline is criteria construction and these criteria are then inputs for situation assessment and coaching decision models. User modelling and profiling approaches are used within the criteria construction step. Once all required criteria for a specific coaching pipeline are determined, the focus will be on defining how these criteria are computed from the available data, either directly (e.g. ambient temperature), or indirectly with some predefined function (e.g. speed of movement) or with a learned predictive model (e.g. detecting specific activities to obtain activity profile). This conceptual model enables us to integrate arbitrary sensing devices and arbitrary modelling and profiling algorithms.

The SAAM platform have been successfully set-up and consists of different sub-systems to be refined once the functionalities, the data protocols and interoperability of hardware for each domain are finalized. New methods were being researched and best practices applied with the creation of the base management architecture and functionality of the system. Several additional services that would allow social integration between senior and their social circles, and provide appropriate monitoring were being initiated (e.g. calendar for personal and social events with information, fed from both different groups of social circles and from seniors themselves).

In terms of persuasion, concepts were created, embedding behavioral design in AAL. By connecting behaviours and devices across homes, technologically enhanced objects can be used as means of persuasion. Behaviour change goals, therefore, become embedded in daily life objects and habits, which respond to the user, but at the same time carefully nudge him towards his goals and commitments.

The project communication and dissemination has been initiated from the beginning of the project with the initial visual identity and strategy and communication pack. Partners participated in diverse types of international and local events and initiatives to communicate the project. Press releases for media have been sent-out and followed-up. For the reporting period over 10 presentations at international events were held and 10 Articles were published in Conference proceedings for the period where partial SAAM results and research, funded by the project was disseminated.
The project's progress beyond the state of the art is exemplified by three core innovations: an unobtrusive and fully ambient data collection mode that does not require user interaction, the use of cheap common hardware for sensing and coaching so as to allow widespread adoption of the system, and the use of primary user's social circles as part of a multi-modal coaching system. The user-side of the SAAM system consists of a set of affordable off-the-shelf devices connected together with the edge gateway that further connects them to the cloud-based part of the SAAM system. These devices include stationary components such as smart power meter, UWB localization sensors and an integrated ambient sensor that enable completely unobtrusive user activity monitoring. Activity monitoring can be further enhanced with optional wearable sensors or smartphones. Based on the sensing data the cloud-based part of the system performs coaching in four main domains of the user's life: mobility (ability to move), daily activity, sleep quality and social activity. Within each of these domains one or more coaching pipelines continuously assesses user's situation and autonomously decides regarding triggering and rendering of suitable coaching actions. A combination of advanced data fusion, machine learning and decision support modelling is used in each of the coaching pipelines. The main innovation of the system is that the coaching actions are not rendered only directly to the user, but primarily through his or her social circle, thus positively affecting social interactions.