SPARO: Snap-On/Rotate-Off Passive Gripper for Aerial Perching
Jakob Domislovic, Antun Ivanovic, Frano Petric and Matko Orsag
AI summary
Problem
Aerial drones are limited by battery life and the high power cost of hovering, making prolonged missions difficult. Existing perching solutions often require custom hardware, extra sensors, or complex control, limiting their practicality and widespread adoption.
Approach
The authors designed SPARO, a 3D-printed bistable gripper that closes automatically upon impact and opens via the drone's yaw rotation, eliminating the need for onboard gripper actuation or sensing. This mechanism is paired with a nonlinear programming trajectory planner that shapes approach paths from below to ensure consistent, reliable engagement.
Key results
- SPARO weighs 290 g and closes in under 45 ms upon impact
- Fully passive operation requires zero onboard gripper actuation or sensing
- Gripper-aware NLP planner generates full-state SE(3) trajectories with below-approach shaping
- Validated across multiple off-the-shelf quadrotors with reliable perching on diverse surfaces
Why it matters
Provides a low-cost, universally mountable perching solution that extends drone endurance for long-duration missions like CBRNE response without requiring custom platforms or complex control.
Abstract
We present SPARO, a fully passive, impact- activated SnaP-on And Rotate-Offgripper for aerial perching that requires no onboard actuation or sensing for closing or opening. SPARO is 3D-printed from PET-G, weighs 290 g, and closes upon impact within 45 ms using a bistable spring mechanism that can grasp objects up to 15 cm in diameter. The gripper opens through a mechanically coupled yaw motion from the host robot, enabling repeated autonomous perching without any additional actuation. A trajectory optimization framework based on nonlinear programming generates gripper-aware, full-state trajectories in SE(3) that promote approach from underneath the target and satisfy closure constraints. Extensive experiments confirm reliable, repeatable autonomous perching on various surfaces and robots, using off-the-shelf quadrotors, thus validating the system’s simplicity, universality, and reproducibility.