Homography-Based Visual Servoing of a Joint-Constrained Robotic Endoscope

In recent years, homography-based visual servoing (HBVS) under remote-center-of-motion (RCM) constraints has been successfully applied to minimally invasive surgery (MIS), which can ignore explicit image depth information. However, the existing HBVS control scheme fails to consider the joint constraints of the endoscopy robot and requires computing the Jacobian pseudo-inverse. To tackle these challenges, this article proposes an innovative HBVS control scheme that incorporates physical limits, RCM and HBVS constraints into a quadratic programming (QP) formulation. Subsequently, the QP problem can be solved by a gradient-based neural network (GNN). After that, simulations and experiments based on a robotic endoscope coupled with a Franka Emika Panda robot are conducted to validate the superiority of the proposed HBVS control scheme in achieving visual servoing.