Autonomous Navigation Strategy Based on Image Data Weighted Fusion for Flexible Endoscopic Surgical Robot

The traveling direction of surgical robot within complex and tortuous natural lumen of the human body is manually adjusted by doctors, leaving room for improvement in operational efficiency and safety. This paper proposes an autonomous navigation strategy for flexible endoscopic surgical robot based on weighted fusion of image data, which enables the robot's terminal orientation to track the lumen center in real-time, avoiding collisions between the endoscope tip and surrounding tissues such as the lumen wall. Firstly, the center position of the lumen is calculated using lumen contours, image intensity, and optical flow vectors. Subsequently, various weighted fusion criteria are designed based on the distances between lumen centers calculated by different methods. Additionally, a turning prediction method is introduced to anticipate the turning direction in advance during the robot's progression. Finally, the proposed autonomous navigation strategy has been successfully tested in a series of straight-line and turning experiments, accurately calculating the lumen center position.