Robotics and AI for Sewer Pipe Inspection and Maintenance

ongoing

There are estimated to be 3.2 million km of sewers in Europe. Their inspection technology is based on human analysis of CCTV images. Often condition of sewers is just not known. This leads to failures such as property flooding and wastewater spills via storm sewer overflows into rivers. PIPEON will develop robotic and AI technologies for autonomous sewer inspection and maintenance. Robots are rugged-by-design working in the corrosive and abrasive sewer environment. They move over different surfaces, above and within the wastewater. Dexterous manipulators install flow monitors and remove blockages of different sizes made from materials such as textiles, fats and grease. Sensors, sensor fusion and sensor perception ML algorithms for feature and anomaly detection facilitate navigation and mapping in self-similar/featureless sewers and risk-aware mission planning methods reduce the risk of robot failure. All these tools will be developed with computationally bounded resources so that robots are able to operate for days. The use of robots automatically determining defect type and size will reduce the cost or inspecting sewer networks by at least one order of magnitude if not more. The reduction in cost “floods” water utilities with data and allows delivers the right maintenance at the right time leading to almost “defect-free” sewer operation. Removing blockages quickly will reduce sewage spills into rivers by an estimated 30% of current numbers. This will intelligently enhance protection of Europe’s rivers and meet Europe’s zero-pollution ambitions in a sustainable, climate friendly way. The team includes scientists from leading European research groups in robotics, in-pipe sensing, mapping, manipulation and the modelling and management of sewer networks. Leading SMEs in the areas of AI based sewer data analysis, and robotics bring knowledge of exploitation pathways. Three water utilities bring case study networks and knowledge of practical application.