Background
Robot-assisted laparoscopic abdominal surgery is widely adopted due to its benefits, such as reduced post-operative recovery time, lower complication rates, and improved patient outcomes [1]. However, precise instrument navigation and intraoperative visualization remain major challenges due to the lack of direct depth perception, tissue deformation, and complex anatomical structures [2],[3]. Traditional robotic-assisted surgery relies on external tracking systems or stereo vision, but these approaches introduce limitations in terms of flexibility, real-time adaptability, and procedural workflow [4].
Aim
ROADIME-ABD seeks to enhance robotic-assisted laparoscopic surgery by integrating Visual SLAM with AI-driven perception for real-time laparoscope tracking and adaptive 3D scene reconstruction. The project aims to develop a robotic-aware intraoperative navigation framework, allowing autonomous laparoscope localization, real-time depth estimation, and adaptive surgical scene reconstruction to improve intraoperative decision-making and procedural safety.
This research is a collaboration between the Robotics and Mechatronics (RAM) group at the University of Twente and clinical partners at The Netherlands Cancer Institute, bringing together expertise in robotic perception, computer vision, and AI for surgical navigation.
Key Research Directions
ROADIME-ABD advances several research fronts in robotic-assisted laparoscopic workflows, focusing on:
1. AI-Driven Robotic Navigation – VSLAM Framework
- Autonomous laparoscope tracking using AI-driven robotic perception and SLAM for precise spatial awareness in robotic-assisted surgery.
- Scale-aware monocular depth estimation to improve depth perception from laparoscopic video, aiding autonomous robotic control and surgical guidance.
- Learning-based motion prediction for robotic laparoscope stabilization in dynamic surgical environments.
2. High-Fidelity 3D Reconstruction
- Real-time 3D reconstruction of deformable abdominal tissue for enhanced intraoperative visualization and robotic-assisted guidance.
- Deformation-aware reconstruction techniques integrating AI-based models with robotic sensing to adapt to tissue movements and surgical interactions.
3. Robotic Integration and Clinical Deployment
- Robotic-assisted laparoscope control strategies, leveraging AI-guided scene understanding for improved maneuverability and stability.
- Experimental validation on robotic surgical platforms and real-world laparoscopic datasets to evaluate system feasibility.
- Real-time deployment for intraoperative decision support, integrating robotic vision with AI-driven navigation models.
By bridging robotic perception, AI-driven image-guided navigation, and real-time intraoperative mapping, ROADIME-ABD aims to push the boundaries of intelligent robotic assistance in minimally invasive surgery.
If you are a student interested in contributing to this project, please contact f.j.siepel@utwente.nl or m.khan@utwente.nl.
References
[1] Kawka, Michal, et al. "Laparoscopic versus robotic abdominal and pelvic surgery: a systematic review of randomised controlled trials" Surgical Endoscopy 37.9 (2023): 6672-6681.
[2] Xu, Lisheng, et al. "Information loss challenges in surgical navigation systems: From information fusion to AI-based approaches" Information Fusion 92 (2023): 13-36.
[3] Iftikhar, Muhammad, et al. "Artificial intelligence: revolutionizing robotic surgery" Annals of Medicine and Surgery 86.9 (2024): 9.
[4] Yang, Liangjing and Etsuko, Kobayashi. "Review on vision-based tracking in surgical navigation" IET Cyber-Systems and Robotics 2.3 (2020): 107-121.