APPLICATIONS / RUBSEE PROJECT
RUBSEE European Project
Beyond the sorting application, our A.I. based vision technology can be used to monitor in real time the waste flow processed by waste treatment plants, to foster their operation so that better economic, environmental and regulatory performance can be achieved. This was the seed of our RUBSEE project: the development of a waste stream monitoring system based on AI for which Sadako have had the financial support of the European Union under an SME Phase 2 – Horizon2020 program.
Project started on February 2017 and has successfully ended on September 2019, with two main outcomes: relevant advances in Sadako’s AI technology for waste detection and 3 pilot systems installed and running in 3 different European waste treatment plants. The following video summarizes Sadako trajectory and the RUBSEE work and results:
RUBSEE monitoring system
In a typical industry, the input raw materials are totally controlled, and all undergoing products are fully oversight in their way to become finished goods. However, the input of an urban Waste Treatment Plant or Material Recovery Facility is a highly variable and uncontrolled waste flow. Nowadays plants work without real time automatic information of the mix of materials they are processing or obtaining along the relevant points of their layout. Although there’s a lot of value in the data collection, NIR conventional sensors or other available technologies are too expensive for that.
To address this unmet need and opportunity, we have developed RUBSEE, a disruptive real-time monitoring system that uses advanced Artificial Intelligence and Computer Vision to determine in every moment the composition (kind/quantity) of material present in a number of locations in the plant. It aggregates and presents the information so that in can be easily analyzed and activated, and generates automatic alerts that can help managers and technical team to detect and resolve undesirable events. With the information created by RUBSEE:
- Operators can adjust the parameters of their current equipment on real time and in some cases even readapt the layout design
- In the same way, unexpected technical failures or equipment yield drops can be immediately detected for quick solving
- Historic information and comparative analysis can shed light to plant operation and allow data-driven decisions
The system can identify what’s moving across the lines, providing to waste operators the same type of actionable information that manufacturing and energy companies already employ.
The project development has faced great technological challenges, mainly related to the complexity of the waste flux. RUBSEE targets the whole plant layout, and so it needs to be able to detect and measure in a big diversity of waste streams, more or less crowded and with a very different mix of objects depending on the location inside the plant. Some of the positions addressed have supposed a great defiance in terms of Artificial Intelligence based visual recognition.
In spite of that, Sadako is really encouraged with the project results and looks forward to next steps to complete development and market delivery (via licensing) of the RUBSEE full monitoring system. We are convinced that RUBSEE can make a great contribution to waste treatment plants, making them smarter and more efficient. We would like to specially thank:
- The CE financial support via H2020 – SME Instrument, key for the development of the project.
- Ferrovial Servicios, which plants has hosted the 3 pilot systems and which contribution and feedback in terms of customer needs definition and usability has been highly valuable and most appreciated.
- The Technological Center Ateknea Solutions, that has developed the Remote Monitoring Platform for data aggregation and management and the ADUS (Automatic Database Update System).
- The Research Center Centre de Visió per Computador, which collaborated in the whole AI Data strategy and supported the Data Base creation.
Below images show some of the monitoring devices installed in the 3 pilot sites as well as examples of the platform interface.