Germany has launched its own Coronavirus Warning App on Tuesday, June 16, 2020. Released by the Robert Koch Institute (RKI), following development by SAP SE and Deutsche Telekom AG, the new app could prove a key digital component in the fight to control and contain the current coronavirus pandemic. The Fraunhofer-Gesellschaft continues to be closely involved in technical aspects of the project – in particular, key components used to estimate physical distance by means of the exposure notification API (application programming interface).
“Since the end of April, the Fraunhofer-Gesellschaft has been advising the consortium around SAP and Telekom on the important project of developing a German coronavirus warning app under the auspices of the RKI,” explains Fraunhofer President Prof. Reimund Neugebauer. In particular, the Fraunhofer Institute for Integrated Circuits IIS has been assisting with the specific technical challenge of optimizing the use of the exposure notification API for the purpose of estimating physical distance. Researchers have drawn on many years of experience in the development of technology to estimate the distance between moving persons by means of signal transfer between two devices on the basis of the Bluetooth Low Energy (BLE) standard.”
“Fraunhofer IIS advises the project consortium of SAP and T-Systems on specification and use, supports with testing the exposure notification API to the Google und Apple operating systems, and offers guidance for implementation,” says Prof. Albert Heuberger, director of Fraunhofer IIS. “We provide technical support, evaluate technical documentation and assist with the harmonization of various functions and formats. In addition, we conduct comparative tests to check the accuracy of the estimates of physical distance. These tests have identified where improvements can be made, for example, to the API and calibration of antenna; both of these issues have now been addressed or resolved by the OS providers.”
Methodologically correct API testing in simulated everyday situations
For this purpose, tests were carried out in simulated conditions designed to correspond to real-life situations. The API was tested in everyday scenarios such as traveling by public transport, waiting in line, dining in a restaurant and being at a party. Each of these scenarios was defined by the RKI. The results were then discussed with project partners SAP, Deutsche Telekom and the RKI. In addition, the configuration of the app was adjusted in collaboration with the RKI. “These tests combine a dynamic change of location of test persons along with real-time reference estimates of the actual distance between them,” explains Steffen Meyer, head of the Cooperative Systems and Locating Group at Fraunhofer IIS. “The only facility where such tests can be conducted in this form and in a methodologically correct manner is the Nuremberg Test and Application Center L.I.N.K of Fraunhofer IIS.”
BLE can be used to detect other Bluetooth devices in the immediate vicinity. Proximity is a necessary but the only criterion for evaluating the risk of infection. The responsibility for determining this lies with the epidemiological community, including the RKI. Device calibration is now being undertaken by providers or manufacturers. The measurements conducted by the app, and the frequency with which these are made, are on the basis of the technical framework currently provided by Google and Apple – the OS (and device) providers and thereby also providers of the corresponding Bluetooth interface.
Back in April, Fraunhofer placed its work towards an app to fight the coronavirus at the disposal of efforts to produce a solution on the federal level. The idea that the analytic labs responsible for processing PCR tests for COVID-19 should be connected to the app was also first raised by Fraunhofer and has now been adopted in the actual solution. In addition, Fraunhofer has submitted ideas for further possible components such as the voluntary provision of ancillary data to the RKI for future research on the pandemic.
This article was first published on 16 June by Fraunhofer.