Tiny flying robots that flap their wings like birds and insects could one day inspect damaged buildings, explore tight spaces and assist in search-and-rescue missions.
However, one of the biggest challenges has been keeping these lightweight machines stable when they are disturbed by wind or sudden air movements.
Now, researchers from Chiba University in Japan have developed a new control system that helps these bird-inspired robots stay much steadier during flight.
Their study, published in Control Engineering Practice, could help make flapping-wing robots more reliable for real-world tasks.
Unlike conventional drones that fly using spinning propellers, flapping-wing micro aerial vehicles (FW-MAVs) generate lift by rapidly flapping their wings.
This gives them impressive agility and allows them to hover like hummingbirds or maneuver more like dragonflies.
Their small size also makes them better suited for flying in narrow or crowded spaces where larger drones may struggle.
Despite these advantages, controlling flapping-wing robots is much more complicated than controlling standard drones. Because they are so light, they are easily pushed off course by wind gusts or other disturbances. Even small changes in the surrounding air can make stable flight difficult.
To better understand the problem, Assistant Professor Abner Asignacion and Dr. Satoshi Suzuki studied the flight behavior of a commercially available flapping-wing robot called the Flapping Nimble+. The robot weighs only about 103 grams, or roughly the same as a bar of soap.
During the experiments, the researchers instructed the robot to hover while moving back and forth at different speeds. They carefully measured how it responded to each command and discovered an unexpected behavior.
Instead of immediately moving in the intended direction, the robot briefly shifted in the opposite direction before correcting itself. Engineers call this “non-minimum-phase behavior.” Although the movement is very small, it makes controlling the robot much more difficult because reacting too aggressively can cause the robot to wobble or even become unstable.
Using this new understanding, the team designed a smarter control system known as a disturbance observer. This type of software constantly estimates outside forces, such as wind, and adjusts the robot’s flight to compensate for them.
The researchers found that the speed of the control system was critical. If it reacted too slowly, the robot remained stable but did not correct disturbances very well. If it reacted too quickly, the robot became unstable and started to oscillate. An intermediate response speed produced the best results, allowing the robot to recover from disturbances while maintaining smooth, controlled flight.
When the new control method was tested, the robot’s position error along its most difficult direction of movement was reduced by more than 53%. Overall, its three-dimensional position error dropped by about 28%, showing a significant improvement in flight accuracy.
The researchers believe these advances could help flapping-wing robots become practical tools in places where traditional drones are less suitable. Because they have no exposed spinning propellers, they may be safer to use around people or in confined environments.
In the future, these agile flying robots could inspect bridges, tunnels and industrial equipment, monitor the environment, or search damaged buildings after disasters.
By improving their ability to stay stable even in challenging conditions, the new control system brings these nature-inspired robots one step closer to everyday real-world use.
