Integration of Visual SLAM in Robot-Assisted Minimally Invasive Surgery: Advances, Challenges, and Solutions

Robot-assisted surgery (RAS) has demonstrated notable advancements in visualization, instrument dexterity, ergonomic improvements, and decreased infection risks when compared to conventional surgical methods. However, within minimally invasive surgery (MIS) contexts, RAS encounters notable challenges in navigating surgical tools effectively. Recent advancements in robot navigation techniques have transitioned from rudimentary wheel odometry and dead reckoning to sophisticated Visual SLAM (Simultaneous Localization and Mapping) methods, capable of addressing complex indoor and outdoor environments. Nevertheless, the integration of Visual SLAM within Robot-Assisted Minimally Invasive Surgery (RAMIS) applications remains substantially restricted due to various factors, including limited field of view, challenges in stereopsis, soft tissue deformations, insufflation effects, and instrument occlusions. This study provides an extensive overview of ongoing efforts towards the development of Visual SLAM algorithms tailored for establishing precise RAMIS systems. It delves into the encountered challenges, delineates the essential features required for establishing a precise Visual SLAM-driven RAMIS system, and explores a diverse range of approaches, which can potentially enhance Visual SLAM functionality within RAMIS contexts.