Autonomous Suturing Framework and Quantification Using a Cable-Driven Surgical Robot

Suturing is required in almost all surgeries but it is challenging to perform with surgical robots due to limited vision and/or haptic feedback. To tackle this problem, we present an autonomous suturing framework that encompasses a novel needle path planner, as well as an accurate needle pose estimator and a six degrees-of-freedom controller. A novel needle grasper is developed that enables needle pose estimation both inside and outside the tissue. The framework was evaluated experimentally using the Raven IV surgical system and important suture parameters were quantified. The experiment results confirmed a needle pose estimation accuracy of < 0.87 mm in position and < 3.46. in orientation across all directions. Moreover, the results revealed that using the proposed framework enabled following the reference needle trajectories with errors of 2.07 mm in position and 4.29. in orientation. These are drastic improvements of more than 10x in position and 5x in orientation compared to the Raven IV kinematic controller. Additionally, the results verified that our framework delivered the desired clinical suture parameters successfully across tissue phantom environments with different mechanical properties and under various needle trajectories. A supplementary video can be found at: http://bionics.seas.ucla.edu/ research/surgeryproject18. htmlhttp://bionics.seas.ucla.edu/research/ surgeryproject18

Automated summary

This study presents an autonomous suturing framework that utilizes a novel needle path planner, accurate needle pose estimator, and six degrees-of-freedom controller to achieve high accuracy in needle pose estimation across all directions. The proposed framework drastically improves suture performance compared to traditional methods, successfully delivering desired clinical parameters across tissue phantom environments with different mechanical properties and needle trajectories.