A bio-inspired methodology of automatic perching for unmanned aerial vehicles

Abstract

Current unmanned aerial vehicles (UAVs) have to stay airborne during surveillance missions, decreasing their energy efficiency dramatically and therefore limiting their endurance significantly. On the other hand, birds perch to conserve energy while maintaining surveillance over their domain. A biomimetic methodology of automatic perching for UAVs is thus a promising solution to their endurance problem.

Firstly, an experimental study of parrots’ perching is conducted to obtain bio-inspirations of perching principles, and the perching procedure is generalized into three stages following which the perching methodology is addressed. Secondly, varying tau-dot, as observed from the approaching flight of parrots, is proposed with a fuzzy logic for perching flight guidance of UAVs and it outperforms the conventional assumption of constant tau-dot in terms of flight time. Thirdly, a scale-dependent expansion model (SEM) is derived for visual perception of tau-dot, and three visual identification algorithms are evaluated for best perception performance. Experiment results verify the effectiveness of the SEM, making onboard autonomous guidance possible. However, improvement on perception accuracy and reliability is still needed. Fourthly, a two-dimensional perching model of quadrotors covering dynamic interaction with perch is established based on analysis of the balancing procedure of parrots after touchdown. Simulation validates the applicability of the model to biomimetic perching. Finally, a gripping perching mechanism featuring force amplification and sensing is designed, and a fuzzy control law is proposed for automatic gripping. Experiments of indoor and outdoor remotely controlled perching, indoor automatic perching, and dynamic automatic gripping are performed, and results show that the perching mechanism is capable of fulfilling reliable and automatic attachment to perch.

The proposed methodology of automatic perching for UAVs covers the complete perching procedure, although its effectiveness has only been verified for each perching stage individually. Future work on enhancement of each component of the methodology and their integration can be done to validate overall effectiveness.