Imagine a bird, with it’s wings flapping, imagine it gliding using the air current the it smoothly descends and perches on a pole. But this is not just any bird, it’s a robot bird. This is a view from the future where robots like these could be used to respond to emergencies or to hunt down drones posing a threat to safety or security.
Uncrewed aerial vehicles (UAVs) are set to form a growing multi-billion dollar industry in the skies with rising potential for a wide range of uses from emergency assistance to drug delivery, home deliveries and precision agriculture.
Professor Anibal Ollero, an electrical engineer at the University of Seville in Spain, says that standard drones with propellers can currently fly for about 20 to 30 minutes, but that the bird-like designs he is working on have the potential to at least double this. “Conventional multirotors are very constrained in terms of time of flight and range,” he said. “We want to increase this range by using the wind and the airflows.”
The GRIFFIN project he currently oversees is seeking to create prototypes of highly autonomous, ultra-lightweight robot birds that can minimise energy in flight, perch on curved surfaces, carry out tasks with moving limbs and artificial beaks, and intelligently interact with people and the environment.
The birds can also carry integrated on-board computers and cameras for visual navigation. “Birds have a very complex body and exhibit complex behaviors. What we are using is the inspiration in the birds to extract relevant features for our robotic ornithopters—aircraft that fly by flapping their wings,” said Prof. Ollero.
To conserve energy, one of the researchers’ aims is to use the wind to enable the robots to soar like birds, combining that with flapping when needed.
Apart from potentially boosting efficiency, Prof. Ollero says a bird-like design has advantages like the ability to reduce noise and improve safety in certain situations because of the absence of propellers and the use of soft, flexible materials like microfibre composites and nylon for the wings and tail.
This means they could be used, for example, to land on injured people and take biometric measurements or put a mask on them in a hazardous environment. “Propellers are not good for this interaction—they can hurt people,” said Prof. Ollero.
Another application is in so-called ‘contact inspection’ in industrial plants, in areas where it may be dangerous for people to work because of the presence of gases. There, a robot could land on a pipe and measure its thickness to detect corrosion.
So far, the team has demonstrated flight with flapping wings both indoors and outdoors, and the ability to land on a small square platform 20 to 30 centimeters wide. The next challenge, which Prof. Ollero says is ‘not trivial,” is to get the birds to perch on curved surfaces such as poles or cables without overbalancing, and then coordinate other functions.
“What we want to demonstrate is these combined capabilities: being able to fly while saving energy, being able to perch, and being able to manipulate their limbs like a bird.”
Challenges faced by the project includes perching, as well as grasping, which the team has been working with materials called shape-memory alloys to create bio-inspired claws that can deform to wrap around a pole—something Prof. Ollero thinks will be possible later this year when combined with machine learning. He also wants to add on-board intelligent sensing of wind to better harness air currents.
Apart from the complexity of coordinating functions, there are some other big challenges to overcome. For one, he says, transitioning between flapping and gliding is difficult, while depending on environmental fluctuations is not always reliable. As lightweight birds with on-board equipment, there are also challenges with the load they can carry—though Prof. Ollero says they had enough capacity to transport things like medicines.
“There is a lot of work to do on the integration of new technologies related to material science, mechanics, aerodynamics and artificial intelligence in our robotic birds,” he added. However, he thinks such birds could be performing practical tasks by 2030.
Visual navigation is another challenge for robot birds, something the project HawkEye is investigating. Its studies into how vision guides flight in birds could potentially translate into more efficient guidance systems for UAVs.
Until recently, birds’ in-flight vision had been little studied, says Professor Graham Taylor, a mathematical biologist at the University of Oxford, UK, who leads HawkEye. “It’s just very hard to study in-flight. That’s challenging until you’ve got miniaturized kit.”
Despite complex bird behaviors on the surface, there seemed to be a relatively simple underlying mathematical guidance law that captures most of the behaviors HawkEye has analyzed, potentially simplifying the creation of algorithms in UAVs that can then be tweaked for different purposes.
“It gives us a common framework for understanding a whole range of behaviours that might not initially seem closely connected,” he said. “Now that we understand the algorithms, the next step is to start applying those on drones.”
The algorithms could be applied to studying other complex behaviours in nature, says Prof. Taylor, while better knowledge on birds and their vision can help understand how they evolved to deal with different tasks, and help design structures such as buildings and wind turbines with cues to prevent birds from crashing into them.
Enrico Ajanic, a Ph.D. student in robotics at EPFL (the École Polytechnique Fédérale de Lausanne) in Switzerland, who has been working on raptor-inspired drones with artificial feathers, says the different types of drones have different advantages at present—with multirotors being agile but with comparatively low endurance, while winged drones currently tend to have high endurance but low agility.
“There are many lessons we can learn from birds,” he added. “They are highly skilled fliers that can negotiate many different flight regimes—fast, soaring, agile and slow flight, to name a few.”
Ajanic said that apart from raptors, researchers can learn a lot from many different bird species. For example, despite their bad reputation, pigeons are also an excellent model for investigation because they can both reliably navigate in cities and fly long distances efficiently, he says.
But he says there are still many challenges, such as the sensitivity of bird-like drones to wind. There are also many questions to answer to improve both the biological understanding of birds and how this can be applied in drones, including how birds twist their tails when gliding and how to optimize the ability of wings to morph.
“There is still a lot of work to do before we can even approximate the flight abilities of birds,” added Professor Dario Floreano, director of the EPFL Laboratory of Intelligent Systems and Ajanic’s Ph.D. advisor. “Birds do not only fly, but also coordinate with each other, interact and change the environment, and make complex autonomous decisions that our drones can hardly match.”
Yet though he says that drones can generate anxiety and risk of privacy violations, he thinks there are more positive potential impacts than negatives—with bird-inspired UAVs having the opportunity to significantly shape the future of drones.
“Such future drones will play an important role in aiding humanity in search and rescue missions, disaster mitigation or efficient delivery,” said Prof. Floreano. “To successfully complete those missions, they will have to navigate autonomously in complex environments. Thus, the challenges such drones must overcome are very similar to the challenges birds are tackling every day.”