A team at Denmark’s DTU is designing a new kind of modular, subsea-resident autonomous robot. Andrew Dykes speaks with one of its architects
ROVs may be one of the offshore industry’s most ubiquitous tools, but that does not mean they are the easiest to use. Operators require considerable skill just to perform what may be considered “simple” tasks – e.g. tightening a valve or bolt on a subsea manifold. Even many smaller models demand the patience, skill and practice that come from an experienced ROV service provider.
But as ROV manufacturers look to new markets, in oil and gas and beyond, many businesses wish to perform their own regular inspections with small, affordable and versatile vehicles. Others would prefer to harness the power of autonomous systems to minimise human intervention altogether. As such, there is a need to push new designs and tools to offer easier operation and greater capabilities.
In response, the Technical University of Denmark (DTU) has opened a laboratory where researchers will design a new kind of modular underwater robot. Central to its creation is the idea that the system should be accessible to a multitude of end-users. “We have a vision of creating a modular robot that is simple to operate – without the need for technical experts. Initially, it will be able to inspect and monitor, for example, the condition of the foundation of a wind turbine under water – or an ocean fish farming plant,” DTU Electrical Engineering Associate Professor Roberto Galeazzi explained.
Dubbed REMORA – REconfigurable MOdular Robotic system for Aquatic environments – the underwater robot will be designed for inspection, maintenance and repair (IMR) of marine structures and environments in confined waters, i.e. the surroundings of larger offshore and subsea infrastructure. But DTU has even grander plans for the modular system. “In the long term, the intention is for the modular robot – which comprises several independent robots (modules) that can both work coupled together and individually – to also carry out repairs on the foundation,” Galeazzi said.
Lessons from Siri The original impetus for development came in 2013 when Galeazzi and his colleagues David Johan Christensen, Jens Christian Andersen and Mogens Blanke were approached by a subsea technology consultant working for the oil and gas sector to discuss some of the challenges related to the use of large work-class ROVs for subsea inspection of oil and gas platforms. This was spurred during the shutdown of DONG Energy’s Siri platform owing to the presence of cracks in several areas of the project’s subsea tank – a failure which resulted in considerable lost production time.
“The major challenges put forward were that current ROV technology was unfit to navigate and operate in narrow spaces like those in focus during the repair of the Siri platform, and the lack of robotic technology that could be utilised to e.g. inspect the subsea tanks from the inside,” Galeazzi told InnovOil by email. “In these challenges we identified an opportunity for disruptive innovation.”
The group outlined parameters which would become the future blueprint for REMORA: that ROVs would have to be autonomous, modular, fault-tolerant, multi-agent (including models for swimming, walking, crawling) and collaborative if they were truly to revolutionise resident subsea IMR. That vision supported two PhD projects in 2014, overseen by Blanke and Christensen, and with further development and private funding in 2015 from the Danish Orients Fond, the REMORA lab was established in February 2016.
Overseen by the DTU Maritime centre, the project involves some of Denmark’s most talented robot researchers, and work in the laboratory includes developing both hardware and software to manage a modular system. Currently the laboratory is equipped with 8 ROV platforms produced by Blue Robotics, and each REMORA unit has a standard package for navigation, including an IMU (3-axis accelerometers, 3-axis gyroscopes and 3-axis magnetometer) and a depth sensor – this is the minimum payload that would be integrated into each module.
Over the past 18 months or so, many students in the REMORA lab have already completed their first prototypes. “The core idea of the modular system is the creation of robots that can dynamically change their morphology by interconnecting with other robots or tools (sensors and actuators) necessary for carrying out a specific task in an IM mission… Working in unison, the robots can draw on each other’s functions, thereby becoming fully autonomous,” Galeazzi said. “For example, thruster failure in a swimming robot will no longer represent a reason to abort a mission, since a fellow agent could swim in, connect with the failing agent and continue the mission together.” The same technique can be applied to other components or missions, also to improve battery performance, for example.
The intention is also for the REMORA system to be subsea resident, using a permanently installed docking station for battery recharging, data transfer and self-repair. The modular robot could therefore install and replace sensors from a subsea docking station placed on the foundations of a wind turbine or drilling platform, providing continuous monitoring.
Achieving this requires substantial engineering work, from software control systems to tools, in order to support a suite of ROV modules that are essentially plug-and-play. Yet the researchers are confident that, when fully developed, these will allow for greater adaptation of robots to individual missions or tasks. In turn, that should help increase resiliency to faults and failures, and will allow better endurance and longer-term autonomous operations.
Modular morphology The concept combines a few of the ideas that have been floating around the world of subsea robotics. Last year, InnovOil reported on the subCULTron research project, which sought to design and build swarms of autonomous robots that could interact and communicate. Meanwhile, most commercial new robotics ventures – Eelume, for example – are already looking at some degree of modularity in their payload assemblies, allowing operators to specify the equipment needed for missions. Few, however, have sought to combine all these disparate elements. Over the past year, the team has achieved much in the way of early-stage development. In addition to setting up the laboratory and the third-party ROV test systems, the group has successfully designed, manufactured and tested the first prototype of the so-called intelligent robotic reconfiguration module (iR2M). This is the black device mounted in front of the robots (visible in the pictures above), which enables ROV interconnection and docking.
According to Galeazzi, further work on the iR2M will enable the possibility of expanding new ROV payloads. He pointed to collaborative work with the SENTINEL project at DTU Aqua, a cognitive robotic system for underwater monitoring and sampling, which aims to upgrade the robotic payload with a sensor module for measuring spectral fluorescence. Another tie-up with the Underwater Time Of Flight Image Acquisition (UTOFIA) project would see a module host a newly developed time-of-flight vision system.
Naturally, other engineering challenges lie ahead. DTU’s ROVs must withstand the demands of their working environment, including marine fouling, strong currents and potentially poor communication signals. Yet at the same time, the units need to remain small enough for the overall system not to become bulky and difficult to manoeuvre, Galeazzi said.
The team must also continue to engineer a connection mechanism – the iR2M – capable of supporting all its planned modules. “To minimise complexity and maximise flexibility, we would like to develop modules that have a universal interconnection mechanism – both mechanically and electronically. This will enable the robot to change its morphology in a completely arbitrary way without being limited by the fact that it can connect only to certain components, due to its own connectivity layout,” he continued.
Swimming unaided DTU also has its eyes on the advances in robotic autonomy over the next 5-10 years. The team has already developed its own vision-based motion tracking system which supports autonomous swimming in a laboratory environment, and currently allows the closed-loop control of a single agent. “After the summer we will upgrade our vision system to a camera with high frame rate so that we will be able to track and control multiple agents at the same time,” Galeazzi added. This system is used in the test tank, but future sea trials will require the robots to be equipped with an acoustic or optical localisation system, he said.
From there, the lab will follow two distinct lines of development: the first will focus on developing the systems for modularity, while the second will lay the groundwork for a subsea communication system. “Plenty of work will be allocated to the reinforcing and testing of the robotic reconfiguration module, and also to the design of a so-called sensor module,” he continued. “On one hand this will allow the robotic system to dynamically change its sensing payload, and on the other hand allow it the creation of a deployable sensor network to monitor things like the structural integrity of wind turbine foundations or oil rig jack-up structure.”
The communication system too poses difficulties, since it is the key to creating a multi-agent system that can collaborate by sharing information while in operation. However, he is confident that improvements in optical communications – increasingly miniaturised equipment in conjunction with the ability to achieve higher bitrate transmission – will enable an autonomous multi-agent subsea network to be realised – and perhaps even sooner than the marine industry thinks.
All of these open opportunities for further industry involvement. Given the scale of REMORA’s ambition, further collaboration will be essential to realising the technology’s full potential. “I believe that the best way to purposefully innovate is to team up with industrial players that have been active in the offshore industry for decades, since they have the know-how about the limits of current technology and the challenges faced in IMR operations,” Galeazzi remarked.
The sheer scale of opportunity for modular, resident subsea solutions like REMORA means the race for development is already on. Yet Galeazzi and his team believe that their awareness of the particular demands of the offshore industry will enable their technology succeed: “We would like to shape our innovation and development around concrete issues, so that the new technology we bring to the table has a true impact on the offshore industry.”
If REMORA can deliver on its potential, we foresee a substantial impact indeed.