A bio-inspired flapping-wing Robot with high degrees of freedom for agile flight. Leveraging learning-based control, the robot is designed to perform agile maneuvers and adapt to complex environments.
Project Overview
This ongoing project focuses on the development of a high degree-of-freedom (DoF) flapping-wing UAV inspired by bird flight. The UAV is designed to perform agile maneuvers and adapt to complex environments.
Flight Test
This is a priliminary flight test of the flapping-wing UAV to demonstrate the capability of a sustained flight.
Flight test demonstrating sustained flapping-wing flight.
Design
The current design of the flapping-wing UAV incorporates several innovative features to enhance its flight capabilities. It has individual degree-of-freedom (DoF) for each wing for pitching and flapping, which allows for more complex flight patterns. Note that the recent design is partially referenced from Prof. Raphael Zufferey’s paper (Zufferey et al., 2021).
Left: Overview of the flapping-wing UAV. Right: A close-up view of the wings.
Control
We are working on using reinforcement learning to control the flappy-wing robot. This method allows the UAV to perform agile maneuvers and adapt to various flight conditions. This methods has shown great performance in the simulation. For more details on the control strategy, refer to the video demonstration shown below and paper (Cai et al., 2024).
Left: Animation of the turning maneuver. Right: Trajectory plot of the turning motion.
Left: Animation of the flapping-wing robot climbing. Right: Trajectory plot of the climbing motion.
Control Video Demonstration
Video demonstration of the learning-based control on the flapping-wing robot in MuJoCo simulation.
We also conducted experiments to test the performance of the flapping-wing UAV in the real world and use the data to improve the simulation accuracy.
Experiment setup for the flapping-wing UAV. The UAV is mounted on a Force Torque Sensor (FTS) to measure the force and torque on the UAV. The UAV in the picture is an early prototype of the flapping-wing UAV.
Conclusion
The High_DoF Flapping-Wing UAV project showcases the potential of combining innovative design with advanced control strategies to achieve agile and adaptive flight capabilities for flapping-wing robots. The integration of these elements allows the UAV to unlock the potential of flapping-wing robots to perform complex maneuvers and operate effectively in diverse environments!
References
2024
Control, RL
Learning-based Trajectory Tracking for Bird-inspired Flapping-Wing Robots
Jiaze Cai, Vishnu Sangli, Mintae Kim, and 1 more author
Bird-sized flapping-wing robots offer significant potential for agile flight in complex environments, but achieving agile and robust trajectory tracking remains a challenge due to the complex aerodynamics and highly nonlinear dynamics inherent in flapping-wing flight. In this work, a learning-based control approach is introduced to unlock the versatility and adaptiveness of flapping-wing flight. We propose a model-free reinforcement learning (RL)-based framework for a high degree-of-freedom (DoF) bird-inspired flapping-wing robot that allows for multimodal flight and agile trajectory tracking. Stability analysis was performed on the closed-loop system comprising of the flapping-wing system and the RL policy. Additionally, simulation results demonstrate that the RL-based controller can successfully learn complex wing trajectory patterns, achieve stable flight, switch between flight modes spontaneously, and track different trajectories under various aerodynamic conditions.
@article{cai2024learning,title={Learning-based Trajectory Tracking for Bird-inspired Flapping-Wing Robots},author={Cai, Jiaze and Sangli, Vishnu and Kim, Mintae and Sreenath, Koushil},year={2024},eprint={2411.15130},archiveprefix={arXiv},primaryclass={cs.RO},}
2021
Design of the High-Payload Flapping Wing Robot E-Flap
Raphael Zufferey, Jesús Tormo-Barbero, M. Mar Guzmán, and 6 more authors
