For birds trying to escape predators, takeoff is a critical stage of flight. A shallow takeoff angle favors rapid acceleration and greater horizontal distance, while a steep angle helps the bird gain altitude more quickly. These strategies prioritize speed and climbing ability in different ways, but their effects on flight performance and the underlying aerodynamic mechanisms remain important questions for further study.
Experiments with living birds are difficult to control because environmental conditions, individual differences, and initial states can vary. This makes it challenging to compare different takeoff strategies under identical conditions. To investigate the problem more systematically, the research team developed and tested a meter-scale flapping-wing robot that could perform different takeoff strategies under controlled conditions, enabling comparisons of flight trajectories, acceleration, and climb.
The team's flapping-wing robot, Hawk 3.0, has a wingspan of 1.82 meters, a takeoff weight of approximately 955 to 1,105 grams, and a maximum flapping frequency of 5 Hz. Its wing design draws inspiration from bird wings: the main load-bearing region is relatively stiff, while the rear region allows greater flexible deformation, balancing structural support with flapping motion.
A key finding of the study is the flapping–pitch effect. During flapping, wing motion and elastic deformation change both the magnitude and the line of action of aerodynamic forces. These changes generate moments that influence the robot's pitch attitude. By adjusting the mean flap angle, the robot can modify these moments, providing an additional means of pitch control during early takeoff, when airspeed is low and conventional tail control surfaces have limited authority.
To analyze this process, the researchers established a two-way fluid–structure interaction (FSI) simulation framework based on the lattice Boltzmann method (LBM). The framework accounts for the mutual interaction between wing deformation and the surrounding airflow and is used to simulate the entire takeoff process. The team also conducted outdoor flight experiments using three preset mean-flap-angle settings. Three trials were performed for each setting, with initial attitude, center-of-gravity position, release timing, and other test conditions kept as consistent as possible.
The experiments showed that different takeoff strategies change how the robot allocates kinetic and potential energy during flight. The high-angle strategy climbed faster during approximately the first two seconds and accumulated gravitational potential energy more quickly. The low-angle strategy accelerated more rapidly, building a greater speed advantage for the subsequent flight.
The researchers used three seconds after takeoff as a common engineering evaluation point. At that time, the tested strategies had reached comparable ranges of altitude, speed, and mechanical energy. Compared with the high-angle strategy of approximately 65°, the low-angle strategy of approximately 15° increased the robot's total flight distance by 49% and horizontal flight distance by 123%.
The results reflect differences in when each strategy prioritizes energy gain. The low-angle strategy emphasizes early acceleration, allowing the robot to build up speed before converting part of its kinetic energy into altitude later in the flight. The high-angle strategy prioritizes climbing and gains altitude earlier, but accelerates horizontally more slowly. By combining robotic experiments with fluid–structure interaction simulations, the study provides engineering-based evidence for understanding how different takeoff strategies affect flight distance.
By combining repeatable robotic flight experiments with fluid–structure interaction simulations, the study identifies the role of the flapping–pitch effect in low-speed attitude control. The findings may help inform the design and control optimization of bioinspired flapping-wing aircraft.
However, the results do not mean that either high angle or low-angle takeoff is always the better strategy. Real birds also rely on leg-assisted takeoff, active wing morphing, sensory feedback, and species-specific abilities and predator–prey conditions. These factors are not fully reproduced by the robotic platform. In addition, Hawk 3.0 uses a preset mean flap angle for each flight and does not yet continuously adjust this parameter during flight.
Future work could explore lightweight servo-driven linkages or adjustable wing-mount mechanisms to enable more flexible in-flight adjustment of wing motion. Further studies could also examine how control strategies perform under different speeds, payloads, and flight tasks. The research offers a testable physical explanation for selected bird takeoff behaviors while opening new possibilities for the design and low-speed control of meter-scale bioinspired aircraft.
Authors of the paper include Haoyu Wang, Siping Zhong, Yihang Shi, Zhenkun Gong, Jizhou Jiang, Zihao Wei, Erzhen Pan, Junzhi Yu, Wenfu Xu.
This study was supported by National Natural Science Foundation of China Grant 62233001 (J.Y.); National Natural Science Foundation of China 62573160 (W.X.); Shenzhen Excellent Scientific and Technological Innovation Talent Training Project Grant RCJC20200714114436040 (W.X.); The Basic Research Program of Shenzhen Grant JCYJ20241202123724032 (E.P.); and Program of Shenzhen Peacock Innovation Team Grant KQTD20210811090146075 (W.X.).
The paper, "Birdlike Takeoff Strategies Enabled by Flapping–Pitch Effect in a Meter-Scale Flapping-Wing Robot", was published in the journal Cyborg and Bionic Systems on Oct 16, 2026, at https://doi.org/10.34133/cbsystems.0714 .
