No abstract

Data and sensor fusion can enable clinical healthcare systems to improve conditions of a patient. However, hospitals are not the only application field of connected medical devices. Domestic monitoring gets more important day by day and applies Internet of Things with mobile sensors, like wearables. Through data processing data is transferred to smart data and personalized recommendations are improvable, if sensors can be chosen individually. Therefore, we developed a generic medical sensor framework which is able to merge any needed sensor and collect data to improve personalized health of an individual. To evaluate our framework and to prove the added value of sensor fusion we present a sensor-based stress detection game.

Background In Palliative Care, actors from different professional backgrounds work together and exchange case-specific and expert knowledge and information. Since Palliative Care is traditionally distant from digitalization due to its holistically person-centered approach, there is a lack of suitable concepts enabling digitalization regarding multi-professional team processes. Yet, a digitalised information and collaboration environment geared to the requirements of palliative care and the needs of the members of the multi-professional team might facilitate communication and collaboration processes and improve information and knowledge flows. Taking this chance, the presented three-year project, PALLADiUM, aims to improve the effectiveness of Palliative Care teams by jointly sharing available inter-subjective knowledge and orientation-giving as well as action-guiding practical knowledge. Thus, PALLADiUM will explore the potentials and limitations of digitally supported communication and collaboration solutions. Methods PALLADiUM follows an open and iterative mixed methods approach. First, ethnographic methods – participant observations, interviews, and focus groups – aim to explore knowledge and information flow in investigating Palliative Care units as well as the requirements and barriers to digitalization. Second, to extend this body, the analysis of the historical hospital data provides quantitative insights. Condensing all findings results in a to-be work system. Adhering to the work systems transformation method, a technical prototype including artificial intelligence components will enhance the collaborative teamwork in the Palliative Care unit. Discussion PALLADiUM aims to deliver decisive new insights into the preconditions, processes, and success factors of the digitalization of a medical working environment as well as communication and collaboration processes in multi-professional teams. Trial registration The study was registered prospectively at DRKS (Deutsches Register Klinischer Studien) Registration-ID: DRKS0025356 Date of registration: 03.06.21.

Die deutsche Bundesregierung zielt darauf ab, bis 2022 eine Million Elektroautos auf die Straßen zu bringen. Bisher erscheint dieses Ziel jedoch unerreichbar, da zahlreiche Autofahrer auf den Kauf von Elektroautos verzichten. Als Grund nennen sie dabei primär die unzureichend ausgebaute Ladeinfrastruktur, die unter anderem aus der Unrentabilität des Betreibens von Stromtankstellen resultiert. Eine Möglichkeit, die Profitabilität solcher Investments zu steigern, ist, die Verweildauer der Ladekunden zu nutzen, um die Ladevorgänge auf monetär günstige Zeitpunkte zu legen. Intelligente Decision Support Systems können die aggregierte Berücksichtigung aller relevanten Einflussfaktoren unterstützen. Bisherige Lösungsansätze aus dem Bereich Green IS setzen beispielsweise auf die Reinforcement Learning Methode Q-Learning. Aufgrund der geringen Skalierbarkeit ist sie allerdings nicht auf größere Stromtankstellen anwendbar. Um auf diese Herausforderung einzugehen, wird in dieser Arbeit ein Deep Reinforcement Learning Ansatz verfolgt. Die Evaluation in einem Realwelt-Setting zeigt, dass die Profitabilität von Stromtankstellen durch den Einsatz dieses Modells deutlich steigt.