My mentor, the cockroach – Or how we built a novel drone
Have you ever tried to fly a drone? Your first concern was probably to manage to keep it in the air and to avoid crashing into obstacles. Most autonomous drones are also built and programmed to avoid flying into trees or walls, which kind of makes sense at first sight. The problem is, they can’t easily navigate cluttered environments like forests. So, what if you actually want to look behind those tree branches? Ask the cockroach!
A dusky cockroach (Ectobius lapponicus) on a leaf. Image: Wikimedia Commons/Adam Opioła, CC licence
To monitor biodiversity in a certain region, researchers often collect environmental DNA*. Drones can be very useful for this task as they can reach not readily accessible locations such as the top of trees. However, since drones perceive vegetation as an obstacle that should be avoided, sampling beyond the forest canopy is not possible. In our endeavor to design a drone that overcomes this difficulty we took inspiration from cockroaches.
Cockroaches, as other insects for that matter, are experts at navigating their environment. Have you ever watched a cockroach effortlessly crawl through a dense patch of grass? It doesn’t stop to figure out each blade of grass; it just pushes, wiggles, and slides its way through, sometimes rolling its body. This concept inspired the design of our new drone aiming to look behind the tree branches.
Cockroaches don’t need complex systems to map their surroundings; they use their bodies to physically engage with obstacles. In the same way, the shell of our new drone, which is formed like a disc, acts like a sensor. It can "feel" obstacles and decide whether to push through them, slide along their surface, or do both at the same time. Kind of how a cockroach would squeeze through a pile of leaves. Add a low-friction surface, like some insects have, and you get a pretty agile drone.
To view this content, you must accept marketing and third-party cookies.
For the curious: The magic of feedback control
The trick behind this new drone’s abilities lies in a smart control system. Rather than using a model to predict every obstacle it encounters, it adapts as it flies. The drone "senses" the obstacle’s stiffness and adjusts its movement — as if you bump against something and decide whether to keep pushing or glide around it. It’s a simple yet effective way of controlling the drone without requiring a lot of computing power.
This kind of breakthrough is particularly exciting for applications like environmental monitoring, precision agriculture, and even search and rescue missions. Imagine drones flying through dense forests to collect data on endangered species or to help farmers check on crops without needing to clear paths. To improve our design, we are planning on giving the drone a fully protective spherical shell to help it navigate even more difficult environments.
By learning from nature, specifically from insects, drone technology is evolving in exciting and unexpected ways. So next time you see a cockroach zipping under a bush, just think – you might be looking at the inspiration for the future of robotics.
Further insect inspiration
In the past we took inspiration from the biomechanical properties of insects’ wings to develop collision resilient drones1. We developed tiny drones that can pull 40 times their weight2. We also developed drones with insect inspired folding wings3.
1. Insects inspire crash-proof drone. Nature 2017, 543:153.
https://doi.org/10.1038/543153a
2. Miceli C 2018. This tiny wasp-inspired drone can pull 40 times its own weight. Science. doi: 10.1126/science.aav7885
https://www.science.org/content/article/tiny-wasp-inspired-drone-can-pull-40-times-its-own-weight
3. Dufour L et al. 2016. A drone with insect-inspired folding wings. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS): 1576-1581. doi:10.1109/IROS.2016.7759255.
https://ieeexplore.ieee.org/document/7759255
Glossary*
Environmental DNA: DNA originating from cellular material expelled by organisms (hair, skin, feces, etc.) into the environment.
Reference
Aucone E et al. 2024. Synergistic morphology and feedback control for traversal of unknown compliant obstacles with aerial robots. Nature Communications 15: 2646.
The author
