Description of the use case

As close to citizens as an emergency eCall System is one of our pilots about. Considering the whole paradigm of medical devices, Fraunhofer Fokus will work with these systems and accountable Artificial Intelligence.

What is the pilot for?

In the context of the SPATIAL project, Fraunhofer FOKUS aims to design and implement an explainable and accountable emergency call (eCall) functionality based on explainable and accountable AI algorithms. Our goal is to investigate the automated detection of emergencies based on eHealth data collected from ubiquitous eHealth sensors and IoT devices such as smartwatches, smart T-shirts, or chest straps. More specifically, we aim at collecting relevant eHealth sensor data, analyse it by using secure, privacy-preserving, explainable, and accountable AI methods, and then automatically initiate an emergency call in case of detected anomalies. In this scenario, the relevant, already analysed data will also be made available to the emergency call center. In addition, we aim to provide a transparent and understandable explanation of why the AI-based emergency system has detected an emergency and initiated an automated eCall. In this way, the emergency call center agent receiving the eCall will be able to assess the emergency situation more accurately and initiate more appropriate countermeasures. Fraunhofer FOKUS plans to integrate this explainable and accountable eCall functionality into a Next Generation IP-based telecom platform with the capability of integrating rich-media emergency calls – a combination of voice, text, IoT sensors, social network information, and video. Precisely, Fraunhofer FOKUS will extend EMYtest, a prototype of an IP-based next-generation emergency communication system developed in the H2020 project EMYNOS[1] (nExt generation eMergencY commuNicatiOnS).


Why is it useful for FRAUNHOFER FOKUS?

Fraunhofer FOKUS can look upon a track record of research works, concepts, and prototypes in the context of emergency communication systems, e.g., in the context of the previously mentioned EMYNOS project. Within the scope of SPATIAL, we expect to enhance our knowledge in this area further and to develop an AI-based eCall demonstrator that is easily transferable to use case studies in other industrial domains. Furthermore, through the examined use case, we would like to test and evaluate the practicability and applicability of the SPATIAL outcomes in industrial environments. Thereby, we want to identify further application areas and best practices that are highly relevant in the context of smart cities. In addition, we would like to strengthen our knowledge and expertise in applying secure, accountable, explainable, and trustworthy AI methods by participating in the SPATIAL project. Since we identified AI to be a key enabling technology for the technological sovereignty of Europe, this will be of great importance for future use cases, projects, and business cases.

What tangible results are expected from it?

In the scope of the SPATIAL project, the main challenge of the use case analysed by Fraunhofer FOKUS is to define effective, accountable, and privacy-preserving methods for analysing heterogenous eHealth data and automatically initiating an NG112 eCall in case of an emergency. Since eCalls are a very critical aspect and need to be issued at the right moment while at the same time it must be avoided to overload the emergency call centers in the backend, the decision-making AI algorithms must be explainable, transparent, and accountable. Moreover, these explainability and accountability aspects are also a key requirement for the potential certification of AI-based NG112 emergency communication platforms. In addition, the collected and processed healthcare data are to be classified as highly sensitive personal data, which should be protected at all costs and therefore only processed by using privacy-preserving methods. Therefore, the resilience and accountability measures, the guidelines for implementing trustworthy AI solutions, as well as the explainable, accountable, and privacy-preserving AI methods examined in SPATIAL could potentially enable and accelerate the design and implementation of an explainable and accountable AI-based emergency call platform.