A new disturbance-observer control method improves flapping-wing drone stability by compensating for external disturbances, reducing positioning errors for compact aerial robots used in inspection, rescue, monitoring, and infrastructure applications.
Flapping-wing micro aerial vehicles (FW-MAVs) could become more reliable in confined and disturbance-prone environments with a new control approach developed by researchers at Japan’s Chiba University. The technique combines flight-characteristic monitoring with a disturbance observer to detect and compensate for external forces that can destabilise small ornithopter-style drones.
Unlike conventional multirotor drones, flapping-wing aerial vehicles use bird-like wings instead of continuously rotating propellers. Their compact size makes them attractive for applications such as infrastructure inspection, search and rescue, environmental monitoring, and equipment inspection inside buildings. However, their lightweight structures are particularly vulnerable to disturbances such as air movement, which can cause positioning errors and unstable motion.
The key challenge addressed by the researchers is a behaviour known as non-minimum-phase dynamics. When the drone receives a disturbance and attempts to correct its position, its initial response can briefly move in the wrong direction before moving toward the intended position. This makes conventional feedback control more difficult, particularly for small aerial platforms where rapid changes in motion can significantly affect stability.
The proposed disturbance observer adds another layer to the flight-control system. It monitors the vehicle’s behaviour and estimates the effect of external disturbances, allowing the controller to compensate for them rather than responding only after a positioning error has developed.
The researchers tested the approach on a commercially available Flapper Nimble+ ornithopter weighing approximately 103g. They first flew the drone along a straight trajectory to characterise its response and determine how quickly corrective actions should be applied. The resulting flight data were then incorporated into the disturbance-observer control strategy.
Testing showed a 53.1% reduction in positioning error along the X axis. Across three-dimensional space, the overall positioning error was reduced by 28%. These improvements indicate that disturbance-aware control could make flapping-wing platforms more practical where conventional drones face size, manoeuvrability, or access limitations.
For electronics and embedded-control engineers, the work highlights the importance of combining motion sensing, system modelling, feedback algorithms, and real-time control rather than relying solely on mechanical stability. Such architectures could help future micro- and nano-drones maintain accurate positioning while operating in tunnels, shafts, industrial facilities, and other confined environments.
