Estimation of Tool-Tissue Forces in Robot-Assisted Minimally Invasive Surgery Using Neural Networks

Sajeeva Abeywardena,Chris Melhuish,Antonia Tzemanaki,Leonidas Droukas,Sanja Dogramadzi,Qiaodi Yuan,Efi Psomopoulou

doi:10.3389/frobt.2019.00056

Sajeeva Abeywardena, Chris Melhuish + Show 5 more

Open Access

https://doi.org/10.3389/frobt.2019.00056

Copy DOI

Abstract

A new algorithm is proposed to estimate the tool-tissue force interaction in robot-assisted minimally invasive surgery which does not require the use of external force sensing. The proposed method utilizes the current of the motors of the surgical instrument and neural network methods to estimate the force interaction. Offline and online testing is conducted to assess the feasibility of the developed algorithm. Results showed that the developed method has promise in allowing online estimation of tool-tissue force and could thus enable haptic feedback in robotic surgery to be provided.

Highlights

Robot-assisted minimally invasive surgery (RAMIS) has gained popularity in the last two decades through use of the da Vinci master-slave surgical system offering improved vision, precision and patient recovery time compared to traditional MIS (Lanfranco et al, 2004)
Sang et al (2017) modeled the dynamics of a da Vinci robot and, in conjunction with measured motor current, estimated the external force applied at the tip of the surgical tool; while Zhao and Nelson (2015) created a 3 degrees-of-freedom (DOF) surgical grasper prototype with joint dynamics modeled as individual linear 2nd order systems to estimate external forces
The developed algorithm is tested in two manners; first, in an offline manner to test if the prediction system is feasible, and to see if the algorithm can estimate forces accurately in an online manner, such that it could have potential to be used in a force feedback system in RAMIS

Summary

Introduction

Robot-assisted minimally invasive surgery (RAMIS) has gained popularity in the last two decades through use of the da Vinci master-slave surgical system offering improved vision, precision and patient recovery time compared to traditional MIS (Lanfranco et al, 2004). Sang et al (2017) modeled the dynamics of a da Vinci robot and, in conjunction with measured motor current, estimated the external force applied at the tip of the surgical tool; while Zhao and Nelson (2015) created a 3 degrees-of-freedom (DOF) surgical grasper prototype with joint dynamics modeled as individual linear 2nd order systems to estimate external forces (up to 2 N). These methods require some form of modeling and simplification (e.g., neglecting friction) which can affect the estimation accuracy. The complexity of these algorithms may not allow for suitable update rates required for haptic feedback, affecting the systems overall stability and transparency

Objectives

Methods

Results

Conclusion