UCSD's AI lets da Vinci robot pick up needles using suture thread
New method handles occluded needles and avoids pinching tissue during autonomous suturing.
Researchers at UC San Diego have introduced an autonomous suturing framework that reimagines how surgical robots handle dropped or inaccessible needles. Instead of the standard approach of direct grasping—which risks pinching tissue or causing the slippery needle to jump—the system treats the suture thread as a helping hand. By first lifting the thread attached to the needle, the robot can then guide its gripper along the thread to securely pick up the needle, even when it's partially hidden or awkwardly positioned.
The framework, accepted at IROS 2026, covers the full workflow: 3D reconstruction of the thread and surrounding tissue, online selection of safe grasp points on the thread, stable lifting, and bimanual tracking until the needle is secured. Policies explicitly account for visual uncertainty to maintain robustness in messy, real-world surgical settings. Tests on the da Vinci Research Kit showed reliable performance even with tangled threads or when the needle was non-approachable, demonstrating a significant step toward truly autonomous suturing in unstructured environments.
- Uses suture thread as an indirect tool for needle pickup, avoiding direct tissue contact
- Works even when needle is occluded or positioned in an inaccessible area
- Validated on the da Vinci Research Kit with robustness to visual uncertainty and tangled threads
Why It Matters
Enables safer, more reliable autonomous suturing by eliminating the risk of pinching tissue during needle pickup.