Magnetic navigation techniques borrowed from nature could be applied to drones according to a study carried out at the US Air Force Research Laboratory
It’s long been known that many animals – most notably migratory birds, but also some mammals, fish and even insects – navigate by sensing the Earth’s magnetic field.
Now, researchers at the US Air Force Research Laboratory have demonstrated that one of the magnetic navigation techniques thought to be used in the natural world could help enable autonomous vehicles find their way without maps or GPS.
The study, published in the journal Bioinspiration and Biomimetics, used computer modelling to first investigate whether it’s feasible for animals to use rare and/or unique combinations of magnetic properties as a type of navigation waypoint or marker.
“This concept has been put forward before through various experimental work with artificial magnetic fields, and simulation work that examines an animal’s motion in the context of ocean current motion and the magnetic field” explained Dr Brian Taylor, from the US Air Force Research Laboratory.
The group used a software simulation to execute several closed loops around a series of goal locations, in a variety of conditions, and found that this would be a viable navigation technique for animals. “From an engineering perspective, the results show how a simple algorithm with little prior knowledge of its environment can successfully navigate to different specified points,” said Taylor.
He added that the approach may provide a way for engineered systems to autonomously navigate without external positioning aides. “Because the algorithm only has limited prior environment knowledge, a detailed map does not necessarily need to be created or maintained prior to a task or mission, which can save on resources, and is promising for situations where creating the map would be logistically difficult.”
This also means there is no need for a detailed map to be carried on board a vehicle or unmanned platform. This is advantages for unmanned platforms where cost, size, weight, and power are at a premium, and a more detailed map translates into more storage space and processing power.
Dr Taylor concluded: “Along the same lines of considering cost, size, weight, and power, the results show that, under the right circumstances, it may be possible for this type of algorithm to succeed without needing a high measurement frequency. This could ease the computational burden of running the algorithm in a real-time / online setting where resources are limited.”