Tendon-driven Grasper Design for Aerial Robot Perching on Tree Branches

Haichuan Li, Ziang Zhao, Ziniu Wu, Parth Potdar, Long Tran, Ali Tahir Karasahin, Shane Windsor, Stephen G. Burrow, Basaran Bahadir Kocer

Introduction

Monitoring and restoring forest ecosystems is critical for conservation, yet it presents significant logistical challenges for unmanned aerial vehicles (UAVs). Dense canopies and complex terrains make navigation difficult, while the limited battery life of multi-rotor drones restricts mission duration. While perching—staying stationary on a structure—is a known solution to save energy, most existing mechanisms are designed for flat, predictable surfaces. Tree branches are irregular, vary in diameter, and are often obscured by foliage. Furthermore, many active grippers require constant motor power to maintain a grip. This research introduces a bio-inspired, tendon-driven system that allows drones to autonomously locate and securely perch on irregular branches without consuming power during the perching phase.

System Design and Methodology

The proposed system integrates a compliant mechanical grasper with a specialized software stack for autonomous targeting.

Tendon-Driven Mechanism

Inspired by the anatomy of bat claws, the gripper utilizes a "passive-closed" design. The structure features stiff thermoplastic polyurethane (TPU) joints that act as springs, maintaining a naturally closed state. A single Dynamixel servo pulls nylon tendons to open the claw. Once the drone is positioned, the servo releases, and the mechanical tension of the joints secures the branch. This design ensures that zero energy is required to maintain the perched state.

The tendon-driven grasper design, featuring modular fingerprint pads, elastic joints, and servo-actuated tendons.

The gripper is enhanced with modular friction pads, or "fingerprint" modules. These soft, textured pads increase friction and allow the claw to interlock with bark irregularities, improving stability on diverse surfaces.

Vision and Trajectory Planning

To identify suitable perching spots, the system employs a DeepLabV3+ semantic segmentation network with a ResNet-34 backbone. This "raptor-inspired" vision system first identifies a stable tree trunk and then locates the nearest horizontal branch suitable for perching. The vision system was trained on the "Urban street tree dataset," achieving a segmentation performance of 72.15% Intersection over Union (IoU).

Approach maneuvers are managed via minimum-snap trajectory optimization (7th-order polynomial). This ensures a smooth, precise approach to the branch, which is critical for avoiding collisions in cluttered environments.

Results and Discussion

The mechanism demonstrated significant versatility and precision in real-world testing.

Branch Versatility and Payload

The grasper successfully perched on branches ranging from 30 mm to 80 mm in diameter. Static modeling revealed a transition in stabilization physics based on diameter: on smaller branches (30–50 mm), the claw "hangs" via geometry, while on larger branches (up to 110 mm), the squeezing force from the elastic joints becomes the primary stabilizer.

Experimental and theoretical analysis of the grasper's performance across various branch diameters (30mm, 75mm, and 110mm).

Operational Performance

  • Targeting Accuracy: The vision system effectively distinguished trunks from branches, providing precise 3D coordinates for perching.
  • Flight Precision: The quadcopter maintained a mean 3D position tracking error of 0.122 m during its approach, well within the physical tolerance of the compliant gripper.
  • Sensing Optimization: Simulations identified 4 meters as the "Goldilocks zone" for vision—close enough to resolve branches accurately but far enough to maintain a sufficient field of view.

Conclusion

This research presents a viable path toward long-endurance forest monitoring. By leveraging passive mechanical states and deep-learning-based vision, the platform can mimic the energy-efficient behavior of perching animals. Future work will focus on improving visual segmentation in high-clutter environments, conducting large-scale testing across diverse tree species, and exploring more durable materials for the claw tips to withstand repeated long-term deployment.

Summary updated

BibTeX
@inproceedings{li2025tendon,
  author = {Haichuan Li and Ziang Zhao and Ziniu Wu and Parth Potdar and Long Tran and Ali Tahir Karasahin and Shane Windsor and Stephen G. Burrow and Basaran Bahadir Kocer},
  title = {{Tendon-driven Grasper Design for Aerial Robot Perching on Tree Branches}},
  booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
  year = {2025},
  doi = {10.1109/iros60139.2025.11247040},
}
Figure