Toolbox and mEthodology for waTeR based Ai projects – TETRA

An appropriate methodology supports and accompanies their implementation. The tools developed are being evaluated on the basis of two use cases on the Rhine, on the Franco-German border (river monitoring and renaturation measures). Flow monitoring is carried out using continuous video data and sensors. In addition, it is being examined how AI algorithms can operate on these devices (Edge-AI). In the field of renaturation measures, AI is used for sustainability. The methodology accompanying the implementation will be based on the PAISE® methodology that has been developed for the industry. If necessary, PAISE® will be adjusted to harmonise with the toolbox developed and provide a guide for future projects. The toolkit will be freely available in a shared ecosystem. The results can be applied by German and French authorities and SMEs, opening up opportunities for cooperation in the field of AI.

After the end of the project, the participants in this eco-system will be supported by the developed TETRA-methodology, that will both provide information on integrating AI-algorithms into a runtime environment, as well as on how to execute an AI-based project. As the methodology will be based on the already available PAISE process model, the bar is set high: PAISE is developed by the Competence Centre for Artificial Intelligence Engineering (CC-KING) at Fraunhofer IOSB and brings together the industry and AI research. It will be researched how the existing methodology can be applied and where necessary adjustments need to be made. The resulting TETRA methodology will give best-practices, guide and help the reader to avoid common mistakes. As tools will be publicly available as open-source, German and French AI-players can trust on their availability and continue working with these tools in the long term. Furthermore, participation or customisation of these tools to the needs of the SMEs are possible.

As AI-projects in the domain of water can be very heterogeneous, we will evaluate the results against two use cases with dissimilar prerequisites: renaturation of inshore water bodies and river water quality assessment. Renaturation measures require expert knowledge, which is often embodied in textual corpora. Water quality assessment deals with large amounts of structured time-series data. For its evaluation it is complemented with expert knowledge. The achievement rate in the heterogeneous use-cases will measure the efficiency of the toolbox.

In progress

Water is an existential resource with immense value for mankind and the environment and is a driver for the economy and prosperity. To ensure sustainable availability, modern tools for efficient and reliable water monitoring are needed. This is reflected in massive fish die-off the Oder. In order to provide better protection for rivers, the TETRA consortium proposes a project blueprint that will simplify and accelerate the use of artificial intelligence (AI) in the field of water management. The tools provided will save cost and time to AI projects. Authorities, small and medium-sized enterprises (SMEs) are supported in this way to keep rivers resilient and healthy. As part of TETRA, an AI toolbox is being developed specifically for the requirements in the water sector in order to implement common use cases. A suitable methodology supports and accompanies their implementation. The developed tools are evaluated at hand of two use cases on the German-French river Rhine (river monitoring and renaturation measures). Flow monitoring is performed using video streamed data and sensor data. In addition, it is examined how AI algorithms can run on these devices (Edge-AI). In the area of renaturation measures, AI is used for sustainability. The methodology accompanying the implementation will be based on the existing PAISE® process model that was developed for industry. Where necessary, PAISE® will be adjusted to harmonize with the developed toolbox and allow a blueprint for follow-up projects. Both the method and the tools developed will be freely available in a shared ecosystem. The results can be applied by German and French authorities and SMEs, opening up opportunities for cooperation in the field of AI.