"What was scientifically interesting about these projects was that we had to link the autonomy of the drone and the autonomy of the ship by using predictive and cooperative control, where the model of the ship's motion differs significantly in dynamics from that of the drone," said doc. Martin Saska, head of the MRS group at the Department of Cybernetics. "All this had to be done in real conditions where there are wind, waves and so on. Those were the biggest challenges we faced," the researcher noted.
At Orlik, the team showed how autonomous drones can facilitate and refine the movement of a floating object (such as a measuring buoy) towed by a robotic ship. "When a ship on a long rope is pulling an object, it is difficult to ensure its precise movement along the desired trajectory. In the case of cooperative manipulation between the ship and the drone, the high power of the ship can be exploited better by the drone's much more agile movement abilities. By appropriately stacking the forces exerted by such a diverse team of drones, much more precise control of the motion of the towed object can be achieved. As a result, we can avoid obstacles or acquire sensory data in precisely desired locations," explained doc. Saska.
Drones from CTU FEL also search for trash on water surfaces. After they find it, they send information to a ship that arrives and collects the litter in the given place. Which is a time and cost-effective procedure. The cooperating teams of flying and swimming robots can also detect the boundaries of cyanobacteria-infested water bodies with great accuracy.
"The motivation for this research is to optimize drinking water intakes at reservoirs that are affected by cyanobacteria. Information on the exact concentration of cyanobacteria and other pollution obtained from multiple reservoir sites simultaneously allows us to accurately model pollution movement, plan water withdrawals or even interventions to reduce pollution. This is not only true for drinking water sources, but also for example on the recreationally used Orlík reservoir, where the public needs to be regularly informed about the water status and where cyanobacteria are a serious problem," explained doc. Saska.
"On a particular mission, the robotic ship is equipped with an autonomously flying drone that can launch from the ship all by itself, quickly analyse the surface and then return to the ship all by itself and land safely. The ship itself can send it to the locations it needs to scan and, based on the real-time measurements, plan its own route to the locations it needs to, taking samples that will provide more accurate information about the water conditions. And along the way, it can pick up litter or even rescue a drowning person," the scientist added, exaggerating only mildly.
When testing the scientific outputs of CTU FEL at the Orlík reservoir, the team of researchers used their own smaller research models of flying robots and an autonomous catamaran, as well as the resulting products of commercial partners. During the project, the Eurosecur Fly4Future collaborators developed and produced a real-sized autonomous boat for carrying out these missions on water surfaces in the Czech Republic and a special waterproof drone that can land on water in addition to landing on the boat. The team from CTU FEL and both companies collaborated with ČZU to collect and process the data taken by the multispectral camera.
These days, the CTU FEL group is also testing the movement of very fast flying swarms programmed for operations in environments with obstacles without GPS signal and communication. The team of fast-flying cooperative robots can use on-board sensors and artificial intelligence to move just above the terrain, but can also fly through the forest or even fly into a building or cave. Swarms of CTU FEL drones can easily deal with GPS signal or pilot communication failure. They do not need a pilot or GPS to fly, which makes them unique not only in the Czech Republic.