Optimized actuator design for flapping-wing robots: A multi-objective approach to mimic natural flapping dynamics

摘要

An optimized actuator for a flapping-wing robot was developed using detailed geometric and physical models to more closely mimic natural flapping dynamics. The robot’s actuator was reconfigured into a linked mechanism and analyzed through geometric equations. The pseudo-rigid-body model was employed to derive mechanical equations. Dual objectives were set for actuator optimization: minimizing both the maximum transmission angle and the potential energy of the flapping motion, subject to geometric and physical constraints. The optimization utilized the NSGA-II algorithm. Additionally, a virtual prototype with rigid-flexible coupling was created for simulation assessments pre- and post-optimization. Multi-objective optimization led to significant performance gains, including a 35.8% reduction in minimum potential energy, a 45.7% decrease in the standard deviation of the angular velocity, and a 10.0% improvement in the actuator angle’s range of angular variation at a flutter frequency of 4.5 Hz, all compared to a geometry-only baseline. These results suggest that the design provides enhanced stability and better replicates the natural dynamics of flapping flight.