Working in sewers presents a myriad of technological challenges. There are currently no robots that can work without direct human intervention and last for a long time in such featureless, harsh environments.
The ambitious project PIPEON* will develop new robotic and AI-based technologies for mapping, monitoring, and maintaining Europe’s sewer networks using autonomous “thinking” robots and AI-based modelling and analysis tools.
The development and application of such new technologies would have major societal, environmental and economic impact. Instead of repairing in-sewer defects and removing blockages after streets and homes have been flooded with sewage, defects can be quickly identified and repaired and blockages removed when they are still small.
Early, preventative repair and maintenance actions will limit the frequency and
volume of sewage spills from sewer overflows into rivers, a key target of the new Urban Wastewater Treatment Directive, which has been recently approved by the European Commission.
Lots of challenges
The project leader, Professor Maarja Kruusmaa at TalTech, explained that although repair and cleaning are very practical problems, working in sewer environments means that their solution first requires the achievement of fundamentally new research results.
"Robots are mostly used to perform dirty, tedious and dangerous work and are sent to hard-to-reach places instead of people," says Kruusmaa, "but few people have remembered that one of the most complex, dangerous and hard-to-reach environments in the world is right here under our feet when we walk home from work every day."
Working in sewers presents a myriad of technological challenges. "There are currently no robots that can work without direct human intervention and last for a long time in such featureless, harsh environments.”
Thus, we need to develop new types of locomotion mechanisms so that robots are able to move even in wastewater, containing solids and fats.
New control approaches so that robots can navigate in darkness in pipes with few landmarks, physically restricted by ever-changing pipe directions and sizes. Sewer robots need to be smart so that they can survive for days on their own.
Underground, there are usually no communication options or GPS signals and there are no precise maps of sewer networks.
"We will use machine learning algorithms to
navigate underground as well as to identify potential defects. But robots are small, and we cannot use unlimited computing power, as is the case, for example, with large language models like ChatGPT.
“We need to make sure that the robot's own small on-board computer is able to process, learn and decide, and this requires a completely different type of artificial intelligence," Kruusmaa explains.
The project received funding from the highly competitive European Framework Programme on Robotics and Artificial Intelligence.
Kruusmaa commented, "We managed to convince the European Commission that, on the one hand, the project has a very big impact on the environment, the economy and society, and that on the other hand, we are the best team."
The team,
led by TalTech, is multi-disciplinary and has 12 partners, including the Norwegian University of Science and Technology, the University of Sheffield, and several European innovative technology companies, such as an Italian start-up Herobots who are developing novel actuation mechanisms, and also several water utilities.
Simon Tait, Professor of Water Engineering at the University of Sheffield explained that "with over 3M km of sewer in Europe, subject to climate change, new environmental obligations and an ageing workforce, water utilities need radically new approaches to maintaining their service to citizens’ – we believe that autonomous in-sewer robots is an approach that can help meet these challenges."
The researchers' ambition is that by the end of the project, they will have evaluated robot prototypes in several European sewer networks, opening the potential for widespread deployment of robots in sewers in the 2030s.