How to detect soft falls on Android (devices)

At the Knime Berlin summit 2016, Prof. Dr. Dominique Genoud presented a novel way to implement a KNIME workflow that perform machine learning and signal processing on an Android  platform. The use case was to detect soft falls (not from a standing position) using an Android watch. This application has a big impact on how we can detect automatically when elderly people fall from their bed of their chair. This work was originally based on the Master Thesis in Business Administration realized by Vincent Cuendet in 2015 at the HES-SO with the help of the FST (Fédération Suisse pour les Téléthèses), an organization that helps disabled and elderly people to keep their autonomy.