Andrew Dykes explores subCULTron, a pan-European research project established with the aim of developing an autonomous, self-sustaining colony of underwater robots
Despite its relative proximity, we do not know as much about the seabed environment as we might like. Even with the latest technologies, many areas remain difficult to map and monitor in detail – a task made even more complex given the vastness of the territory and the water depth at some of the ocean’s most remote points of interest. It makes sense, then, to turn to robots for a helping hand.
Yet while ROVs and manned submarines are capable of reaching great depths – and sending communications when they get there – they are almost always tethered to surface vessels. Even the best autonomous underwater vehicles (AUVs) have limited range and capabilities, given that they are typically forced to resurface in order to receive position information, transmit data and recharge. This makes searching and scanning the deep-sea floor for specific objects or features difficult and expensive.
The answer, as the subCULTron project has proposed, may lie in a swarm of self-powered, autonomous robots. These machines would source, communicate, store and relay data to the surface as a swarm, enabling the exploration and mapping of large areas with no human control.
aFish called Wander
In the shallow waters of the Venice Lagoon, three alien species are lurking. The first of these, artificial fish or aFish, are agile information-gatherers. These robots can explore and monitor their environment and act as information couriers, relaying data to other parts of the system. Electrodes on the surface of these craft project an electric field around the robot, enabling them to sense information about its surroundings. Such a method is prudent, in part because it enables communication even in shallow and dirty waters.
On the seafloor, artificial mussels (aMussels) are the “collective long-term memory of the system.” After their deployment, these pod-like craft can collect and store data transmitted by the aFish. At the same time, sensor packages can monitor their surroundings over long periods, collecting information on, for example, fish activity or algal growth.
The aMussels’ storage capability is central to the concept, in that it allows data to be kept and transferred beyond the lifetime of individual units. Moreover, the team intend for the units to be manoeuvrable – using water currents, they system could then move and reconfigure itself as a group, if necessary.
The third component, artificial lilypads (aPads) then act as floating communications hubs. With a solar panel topside and docking mechanism below, both aFish and aMussels can recharge at these pads, as well as transmit data. The docking system is actuated, and contains a small water-resistant motor that provides the “grabbing” motion and also keeps firm hold of the robot. The data can then be sent to other units, or back to a topside receiver nearby. The team also says that aPads will be capable of sourcing information from other sources, such as ship traffic or satellite data.
Together, these robots would be controlled by “bio-inspired algorithms.” Mirroring natural functional communities such as bee colonies or flocks of birds, the systems work without a central unit of control but are capable of making decisions collectively. These swarm control algorithms are the work of the Austrian University of Graz Artificial Life Lab – based on work by Lab’s CoCoRo project, an earlier foray into swarm algorithms and which also produced the working design for aFish– and would enable each unit to be aware of the status of every other unit. However, the colony would still be fully operational even if a single unit failed. Current plans envision a fleet consisting of around 150 robots.
This nod towards resilience and disposability is also crucial to the design of the subCULTron system. In large oceans and at great depths, the risk of lost or damaged vehicles is high – prohibitively so in the case of large ROVs, where millions of dollars can be put at risk. Designing and building robots that are cheap enough to be treated as disposable means that subCULTron can be deployed to collect more data in more remote areas with fewer risks. “If you lose 20 % of your swarm robots, it means 80 % of your swarm is still collecting data and is still doing some work,” project co-ordinator Dr Thomas Schmickl of the Artificial Life Lab told the EC’s Horizon Magazine . “If you lose your one field robot, then game over.”
Teach a robot to fish…
Venice, too, is a worthy proving ground for the project. According to the team’s initial plans, the swarm will investigate the channels of Venice itself, as well as the surrounding salt marshes and nearby a mussel farm. In addition to being an ecologically diverse area, it provides practical challenges for the units to navigate – e.g. varying infrastructure, investigating burst pipes and assessing the general impact of transport and tourism on the aquatic environment.
Some problems remain. Wireless communication in particular poses a challenge for researchers – the high-frequency radio signals preferred topside are quickly lost in water. Sonar is useful, but the potential to lose signals among an ocean of noises is high, so other options need to be explored. In addition to the electromagnetic spheres mentioned above, blue LEDs may also allow the robots to transmit information – blue specifically because blue light will travel the furthest in water.
According to an April update from the project team, the docking mechanism between the aMussel and aPad has also been a key focus during the project’s first year. For charging and data transfer to work properly, autonomous docking must be repeatable and reliable, and must be achievable without using vast amounts of energy. Tests this year have succeeded in docking an autonomous aMussel with a radio-controlled aPad, but the team have stated that: “One of the next steps in the second project year will be to make the aPad platform and the docking [fully] autonomous.”
Powering the fleet also raises particular challenges. Although solar may be ideal for the 2m depths of Venice, in deeper waters, where a trip back to surface for solar power may be kilometres, other options will have to be considered. Schmickl has opined that a bacterial fuel cell could be a potential avenue of exploration. This would serve a dual purpose, in that the bacteria digest pollutants in the water to produce energy output, meaning the robots might even clean up the oceans as they go.
The project is still in its early days. The team will work together until April 2019, meaning there should be plenty of time to expand and improve the systems they currently have. But the implications could be significant, and the AUV fleets of the future may owe a great debt to the tiny robots currently exploring the depths of Venice.
Contact: Assoc. Prof. Mag. Dr. Thomas Schmickl Tel: +43 316 380 8759 Email: firstname.lastname@example.org Web: http://www.subcultron.eu/
Could you offer some background on how the Cocoro project has informed SubCULTron?
The aFish robot in subCULTron will be a larger and stronger version of the „Jeff“ robot in CoCoRo. Instead of just 1.5m it will be able to dive 15m deep, but some properties (shape, sensors, …) will be very similar, just larger/more.
How many units are now in the fleet?
Currently we have 5 aPads, 10 aMussels and 0 aFish. At the end it will be 5aPads, 120-130 aMussels and 25 aFish
Has there been commercial interest in the scheme, for use in monitoring or other applications?
Especially the aMussel concept is drawing a lot of attention for long-term monitoring.
Your thoughts on the project in 2016 so far.
I am proud that we came up with a new idea fro a totally new robot type (aMussel) and actually have done it within just a year. And that we have done first autonomous interaction between this new robot (aMussel) and another one (aPad) by autonomous approach, finding, docking, repositioning and undocking.
What next for the team/project?
Further refining and building more robots. going to Venice next spring and test everything (especially the interactions between robots) in Venice. making first long-term observations (weeks) to see how the „fouling“ in the water affects/impairs the robots.