back

2019-09-03

Interview with Dr. Matěj Karásek, Guest Researcher at the Faculty of Airspace Engineering at the Delft University of Technology (TU Delft)

What were the biggest challenges in the development of the DelFly Nimble, compared to their previous versions?

Taking inspiration from birds and insect, the DelFly robots fly by flapping their wings. For control and flight stability, all the previous DelFly versions used a tail, just like the one conventional airplanes have. Winged insects and hummingbirds, on the contrary, stabilize the flight actively by continuously adjusting their wing motion patterns in response to disturbances such as wing gusts. This also allows them to perform quick maneuvers.

The major challenge in the development of DelFly Nimble was to design wing actuation mechanisms that would enable such bioinspired adjustments of the wing motion within the size, weight and power constraints of a 30-g robot. Just like animals move their wings in response to their sensory and neural systems, the robot had to be equipped with miniature motion sensors and a computer. The on-board algorithms estimate the robot’s orientation and give commands to the actuation mechanisms to keep the flight stable.

What are the next steps and applications you would like to accomplish?

DelFly Nimble is one of the most agile flapping wing robots, yet its flight capabilities remain inferior to its natural counterparts. Further research should focus on improving the thrust production, power efficiency and controllability, which goes hand in hand with deeper understanding of the underlying aerodynamic mechanisms and the sensory-motor interactions in natural fliers.

Some applications are close to reality already with the current level of technology. The Nimble could be turned into a toy drone with biological appearance, but swarms of such flying robots, resembling birds or giant insects, could also be used in amusement parks, during art venues and festivals. Because of their light weight and thus inherent safety, DelFly-like robots can be safely operated near humans. We have started a spin-off Flapper Drones, which is developing the flapping-wing technology for commercial applications, with the current focus on applications within the entertainment sector.

In longer term, flapping wing robots have the potential to be further miniaturized and operated in swarms, which requires further research into computationally efficient computer vision and artificial intelligence. Swarms of these intelligent autonomous robots, equipped with sensors, could be used for air pollution monitoring, in warehouses to monitor stock or even for search and rescue applications.

Is it possible to use the DelFly Nimble to pollinate plants? If yes, what features are missing at the moment and how long do you think it would take to develop them?

At the moment, the DelFly Nimble has a very basic on-board intelligence, so in most cases remote operation by a human pilot is necessary. Keeping the actual mechanism of pollination aside, the need of one pilot per “robotic bee” would currently make this solution uneconomical. Before a fully autonomous operation is possible there remain many challenges to be solved, from how to navigate and localize the flower that should be pollinated to the specific mechanism of pollination. Battery autonomy and resilience to all types of atmospheric conditions are just two examples of what also needs to be significantly improved.

However, the ongoing technical advancements in miniature yet powerful computers, sensors and battery technology as well as technologies like 3D printing and MEMS will eventually enable to develop miniature fully-autonomous aerial robotic pollinators. The scalability to the size of winged insects and minimal disturbance of plants when flying over them due to light weight of the robot is one of the benefits of flapping wing technology over traditional designs with propellers.

While in a controlled environment of greenhouses this might be achievable within the next 10 years, the cost of production of each individual „robobee“, together with their limited battery autonomy, will likely limit their real-world use in the near future. In any case, we hope such robotic bees will remain to be our backup solution that will never have to be used.