A collaboration between ULC Robotics and SGN, one of the largest gas network companies in the U.K., has reached the halfway point in a project looking to create an autonomous robotic system that can be used for roadway excavation projects within the utilities sector.
The Robotic Roadworks and Excavation System (RRES) project, which began in early 2018, combines a powerful industrial robot, an all-electric track drive system, below-ground locating sensors, artificial intelligence, machine vision, and new vacuum excavation methods for safer, faster autonomous road work projects. The project, which is funded through U.K. energy regulator Ofgem (Office of Gas and Electricity Markets), has incorporated advanced robotics, artificial intelligence and sustainable components in an effort to improve the efficiency, safety, and lower the environmental impact and cost of utility excavations.
“We are at the midway point of the project and in partnership with SGN we have been able to create the world’s first functional all-electric autonomous excavation robot,” said Ali Asmari, Ph.D., program manager at ULC Robotics. “There is a significant amount of work remaining, including the development of additional tools and support equipment, as well as testing and validating the robotic operation in different environments, but we have an outstanding team and are confident that the robot will be ready to work come 2021.”
The RRES is currently conducting autonomous operations, including cutting of the road surface and performing a patented vacuum excavation method. Initial field trials of RRES are scheduled to take later this year on the SGN network, the companies said.
“Because we operate 76,000 km of underground pipelines in the U.K., and dig thousands of excavations every year, we recognize the need for smarter roadworks,” said John Richardson, head of innovation at SGN. “We’re leading the effort to change utility excavation by investing in the development of technology to address this global problem.”
Richardson said the development, learnings and outputs from the project “have gone above and beyond our expectations. We see RRES as a platform that can expand to meet the needs of the global utility and construction industries.”
The team initially started the project through computer design, simulation and testing, with prototype development and testing in lab scenarios. The project began through Ofgem’s Network Innovation Competition, which agreed to fund the three-year project.
A primary driver for the project is to reduce third-party damage, which happens sometimes when workers damage underground lines owned by the utility, or a separate utility, during the excavation project. Reducing the carbon footprint is another driver, which is why the system is designed to be an electric vehicle, powered by batteries, and wireless. “Right now, big backhoes, dump trucks and other equipment are used during the normal course on maintenance and repair operations for utilities,” said Robert Kodadek, president and CTO at ULC Robotics.
How the system will operate
The idea behind the RRES is that workers will bring the machine on site, and it will drive off the back of a flatbed truck or out of a trailer. Workers will tell it essentially where to excavate, and the system will scan the ground using its sensors to determine whether there’s any buried infrastructure within the first layer of asphalt and concrete. Then the system will cut a hole in the roadway and remove the core created by the cutting. The system will then scan the ground again to see if there’s anything buried within the next layer of dirt.
“We call it soft touch excavation, where it will excavate down to a layer that was determined to be clear through the sensing,” Kodadek said. “It gives us greater surety around our excavation and lets us move as rapidly as possible. So we excavate in layers and scan all the way down to the buried assets.”
The team said while the focus of this project is on natural gas lines, it could eventually be used to excavate down to electrical lines, telecom lines, sewer or water pipes, with the eventual goal of being able to do repair and maintenance tasks on those systems.
In addition, the team stressed that the goal of the system is not to take people out of jobs, but rather provide them with better tools and a safer environment to do this operation effectively. “It’s an operation that doesn’t go away,” Asmari said. “Excavation has to be done as long as we have utilities buried under the ground. It’s just a matter of are we putting our people in danger of getting exposed to these dangerous environments, or creating better tools for them to use.”
Developing integrated sensors
Instead of picking one specific sensor for the scanning task, the project team researched all types of available sensors to determine the best kind for the system. “We called out and brought in every sensor, vendor and manufacturer, every product that had a claim that would be potentially useful in the detection of underground utilities,” Asmari said. “We tested and evaluated all of them and came up with an integration model of not just using one specific sensor, but a suite of sensors based on different technologies, such as electric, magnetic waves, acoustic, and what we’re doing now is integrating the best of these worlds together in one platform.”
In addition, the sensors are integrated with the robotic platform to give the system very accurate localization information. While sensors available on the market can be purchased off the shelf, they end up being deployed manually by operators, and the localization and data processing is less accurate because of potential mistakes made by operators during the data collection and processing.
“The problem is that sensors that might be successful in a lab environment can be less successful on site, because an average operator might not be paying close attention to the details that a robotic system would not miss,” Asmari said.
Advanced data processing is also being developed with the RRES to speed up decision-making time for when and where to excavate. Currently, data collection is done on site and then sent to experts offsite who look at the data and make sense of it, creating a central map. With the RRES, the team is using AI, machine learning and other data processing techniques to look at the data on site and make decisions on site without needing to have an expert readily available or on call. Advances in communications and edge AI technologies are also being explored, Kodadek said.
New ways to excavate
In developing a more efficient method of excavation, the team is using a combination of compressed air and vacuum methods, along with custom tooling that can be controlled as an end effector by the robot to rapidly excavate soil.
“One of the things we’ve done a lot of work on in simulations is to optimize the use of different supersonic nozzles to create the most powerful compressed air blast for breaking up soil,” Kodadek said. “We’ve spent a lot of time doing finite element analysis and simulation, but then also building prototypes of these nozzles, and we feel that we have something that’s really, really special.”
Testing, then deployment
At the end of the project, the team plans to perform live excavation with the system in the U.K. through its collaboration with SGN. Throughout the project, utility companies around the world, including the U.S., have expressed interest in developing particular components of the system for their own projects, Kodadek said.
“There are specific components of the project that may end up being expanded into their networks before the entire system is ready,” he said. “Some of the excavation technology, for example, can be broken out, and some of the development that Ali and his team have done are so relevant to the work [the utilities] do every day, that they’re interested in taking that individual component and developing it further for their applications.”
He added that he hopes people will understand how ground-breaking, pun intended, this project is.
“We’re taking technology that is being used in a factory setting, and putting it out in the field to be able to perform robotics as a service and make the traditional techniques safer, faster, cheaper, and at a lower carbon footprint than what’s been done before,” Kodadek said. “So it would be great if people looked at it and said, ‘Wow, these guys are doing amazing work to the benefit of society.’ “