ISRL - Research

Our Research Question:

What are the key challenges and solutions for automating endoscopic surgery troughout the gastrointestinal (GI) tract, especially in the colon?

Strategic Solution 1: Insertion Assistnace (삽입 보조)

We developed Easycolon, a soft robotic actuator designed to support colonoscopy insertion by straightening the sigmoid colon and generating peristaltic motion. Traditional colonoscopy often faces difficulties due to the curved and flexible shape of the sigmoid colon, which can cause patient discomfort and increase procedure time. Easycolon addresses this by mimicking natural intestinal movements, applying gentle peristaltic waves that guide and straighten the colon. This approach not only reduces insertion resistance and potential risks but also improves safety and comfort for patients. Our work demonstrates how soft robotics can provide practical clinical assistance in challenging anatomical regions, representing a step toward more patient-friendly and efficient endoscopic procedures.

Strategic Solution 2: Lesion Sensing in Confined Spaces (병변 감지)

We developed a soft, inflatable tactile sensor equipped with vision sensing to enable inspection in constrained and confined spaces where conventional sensors cannot operate effectively. The sensor combines the flexibility of soft materials with an internal camera system, allowing it to deform safely while capturing detailed tactile and visual information. This design makes it well-suited for applications in environments with limited accessibility, such as narrow surgical pathways or industrial pipelines. By integrating vision and tactile feedback in a compact, inflatable form, our approach provides a versatile tool for safe and effective inspection in challenging conditions.

Strategic Solution 3: Automatic Calibration (자동 보정)

We developed MonoEndoCal, a deep learning-based calibration framework for tendon-driven flexible endoscopic surgical robots. These robots provide superior access and maneuverability, but precise control is limited by nonlinear hysteresis caused by friction and tendon slack. MonoEndoCal addresses this challenge by estimating bending angles and compensating for hysteresis using only monocular endoscopic images—without additional sensors or markers. Leveraging state-of-the-art vision models, the system performs real-time angle estimation, builds a nonlinear hysteresis model, and applies feedforward control for precise motion compensation.

Strategic Solution 4: Safety-Ensured Automation System (안전성 보장 자동화 시스템)

We developed STITCH, a robotic framework that enhances dexterity for performing suture throws, including the complex tasks of thread coordination and handoffs. Traditional robotic suturing is limited by the difficulty of handling flexible suture threads and coordinating multi-arm maneuvers. STITCH addresses this challenge by integrating advanced motion planning and control strategies that allow robots to manipulate and transfer sutures more naturally and reliably. This approach enables consistent execution of complete suture throws while reducing surgeon workload and improving surgical efficiency. By demonstrating robust thread coordination and automated handoffs, STITCH represents a significant step toward autonomous and precise robotic suturing in minimally invasive surgery.

We proposed a sigmoidal auxiliary tendon-driven mechanism designed to reinforce the structural stiffness of hyper-redundant manipulators used in endoscopic surgery. Flexible and hyper-redundant robots offer excellent accessibility in narrow and curved anatomical pathways, but they often lack sufficient stiffness for precise surgical tasks. Our mechanism introduces an auxiliary tendon path shaped in a sigmoidal form, which significantly enhances stiffness without sacrificing flexibility. This approach improves positional stability during procedures and enables safer, more accurate surgical operations, highlighting its potential as a key technology for next-generation endoscopic surgical robots.

We developed a robotic control framework that enables safe automation of cutting tasks in Endoscopic Submucosal Dissection (ESD). ESD is a complex procedure requiring precise manipulation to remove lesions without damaging surrounding tissue. Our framework integrates safety constraints into the robot’s motion control, ensuring that the cutting tool remains within safe boundaries while maintaining efficiency and accuracy. This approach demonstrates the potential of robotic automation to enhance both safety and reliability in advanced endoscopic surgery.

Strategic Solution 5: System Integration (시스템 통합)

We introduced K-COLON, an endoscopic surgery robot designed to facilitate insertion through the curved colon while ensuring positional stability against external forces. Conventional endoscopic procedures often struggle with the complex anatomy of the colon, leading to difficulties in insertion and maintaining a stable surgical environment. K-COLON addresses these challenges with a novel robotic structure that eases navigation through curved pathways and resists unwanted displacement under external loads. This design improves surgical precision, enhances safety, and provides a stable platform for complex endoscopic interventions.