An expertise-oriented training framework for robotics-assisted surgery

Abstract

This paper proposes an expertise-oriented training platform for robotics-assisted minimally invasive surgery. The framework builds on previous work of the authors and makes use of dual-user teleoperation scenario, allowing the presence of an expert in the training loop. A Fuzzy-Logic (FL) methodology is proposed, which specifies the level/mode of the training required for the trainee according to his/her level of proficiency over the task. A major advantage of the proposed FL approach is that, having the expert in the loop, it can specify the trainee's proficiency level relative to that of the expert in real-time. Moreover, based on the relative skills assessment, the proposed FL approach decides if or to what extent the trainee should receive a haptic guidance force based on Virtual Fixtures or the environment force from the interaction between the surgical instrument and tissue at the slave side. In addition to the level/mode of the haptics-enabled training required for the trainee, the proposed FL framework specifies the authority level of the trainees over the operation in real-time, according to their proficiency levels over the task. Stability of the overall closed-loop teleoperated system is also investigated using the small-gain theorem, resulting in a sufficient condition to guarantee stability in the presence of constant communication delays. Finally, experimental results are given to evaluate the design and feasibility of the proposed framework.