From Ceilings to Walls: Universal Dynamic Perching of Small Aerial Robots on Surfaces with Variable Orientations

TL;DR

This work tackles universal dynamic perching of quadrotors on surfaces with variable orientations and across different sizes. It combines a non-dimensional geometry framework with a Soft Actor-Critic–based policy that ingests $[\tau, \vartheta_x, D_\perp, \theta_{plane}]$ to trigger and execute perching maneuvers, trained in simulation and validated via zero-shot Sim-to-Real experiments. Key findings show that maintaining dimensionless geometric proportions yields similar perching behavior across scales, leg stiffness has limited impact, while damping significantly influences inverted landings and the existence of a perpendicular velocity threshold $V_\perp^{crit}$. The results advance autonomous drone perching capabilities and set the stage for more robust urban-environment operations, with future work focusing on refined damping models, miniature landing gear, and onboard sensing for improved autonomy.

Abstract

This work demonstrates universal dynamic perching capabilities for quadrotors of various sizes and on surfaces with different orientations. By employing a non-dimensionalization framework and deep reinforcement learning, we systematically assessed how robot size and surface orientation affect landing capabilities. We hypothesized that maintaining geometric proportions across different robot scales ensures consistent perching behavior, which was validated in both simulation and experimental tests. Additionally, we investigated the effects of joint stiffness and damping in the landing gear on perching behaviors and performance. While joint stiffness had minimal impact, joint damping ratios influenced landing success under vertical approaching conditions. The study also identified a critical velocity threshold necessary for successful perching, determined by the robot's maneuverability and leg geometry. Overall, this research advances robotic perching capabilities, offering insights into the role of mechanical design and scaling effects, and lays the groundwork for future drone autonomy and operational efficiency in unstructured environments.

Figures (8)

• Figure 1: (a) Image capturing the body-swing stage of a quadrotor during an inverted landing in an experimental test. (b) Corresponding simulation image at the same point in time. (c) Schematic of the dynamic perching policy network. (d) Visual depiction of the dynamic perching sequence, illustrating the progression from flight, through the triggering ($a_{Trg}$) and rotational action ($a_{Rot}$), to the body-swing and final landing stage.
• Figure 2: (a) Depiction of X-Z plane system geometry is projected onto. (b) Source One and Impulse Micro quadrotors with non-dimensionalized Semi-Narrow Short leg configuration
• Figure 3: (a) Gravitational moment contributions: (a1) suboptimal moment; (a2) optimal moment with gravity vector perpendicular to leg vector. (b) Angular momentum transfer: (b1) suboptimal transfer; (b2) optimal transfer with touchdown velocity vector ($\hat{v}_{TD}$) perpendicular to leg vector.
• Figure 4: Experimental setup with an illustrated flight trajectory.
• Figure 5: Polar plots illustrating the landing success rate (color spectrum) for the complete set of approach conditions tested in simulation across five surface orientations [0°, 45°, 90°, 135°, 180°], labeled a-e, and experimentally validated conditions (colored dots). The dashed box indicates the range of conditions testable within the constraints of our experimental setup (due to limited thrust capacity and lab space). Radial lines represent constant flight speeds, while angular lines denote the flight angle relative to the horizon during the quadrotor’s approach. Results are based on the Semi-Narrow Long leg configuration on the Source One quadcopter.