Gaussian Process Regression for Sensorless Grip Force Estimation of Cable-Driven Elongated Surgical Instruments

Abstract

Haptic feedback is a critical but a clinically missing component in robotic Minimally Invasive Surgeries. This paper proposes a Gaussian Process Regression(GPR) based scheme to address the gripping force estimation problem for clinically commonly used elongated cable-driven surgical instruments. Based on the cable-driven mechanism property studies and surgical robotic system properties, four different Gaussian Process Regression filters were designed and analyzed, including: one GPR filter with 2-dimensional inputs, one GPR filter with 3-dimensional inputs, one GPR Unscented Kalman Filter (UKF) with 2-dimensional inputs, and one GPR UKF with 3-dimensional inputs. The four proposed methods were compared with the dynamic model based UKF filter on a 10mm gripper on the Raven-II surgical robot platform. The experimental results demonstrated that the four proposed methods outperformed the dynamic model based method on precision and reliability without parameter tuning. And surprisingly, among the four methods, the simplest GPR Filter with 2-dimensional inputs has the best performance.

Impact of Jaw Orientation on Grip Force Estimation for a da Vinci EndoWrist Surgical Tool

“…This estimation technique compares favorably to existing grip force estimation results reported in literature, even with the non-fixed jaw orientations. As an example, the results in [1] report an average error of 0.07 N for grasps with a peak force of roughly 1 N. This was accomplished via Gaussian Process Regression. Our results, when converted to force, resulted in an RMSE of 0.29 N for grasps with a peak force of approximately 10 N. When comparing the error percentage of peak grasping force the results in [1] yield 7% error, while our results yield 3% error.…”

“…The utlization of accurate grip force estimates during surgical procedures has been widely proposed as a benefit for robotic surgery. Several research publications have proposed varying methods for obtaining this estimate with generally high accuracy [1,2]. Despite these positive results, many of the proposed methods neglect the impact that the jaw orientation (i.e.…”

Surgeon-Centered Analysis of Robot-Assisted Needle Driving Under Different Force Feedback Conditions

“…For example, position exchange is an algorithm that estimates the forces based on the error between the desired position of the robot and the actual (current) position of the robot (Siciliano and Khatib, 2016). Other studies suggested more advanced haptic feedback estimation algorithms for RAMIS (Anooshahpour et al, 2014;Dalvand et al, 2014;Li and Hannaford, 2017;Rivero et al, 2017), as well as advanced algorithms for learning the dynamics of robots that can be modified to cable-driven robots (Rubio, 2012;García-Sánchez et al, 2018;He and Dong, 2018;Rubio et al, 2018;Yen et al, 2018). Each one of the force feedback algorithms has a trade-off between system stability and transparency along with other limitations.…”

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

“…Neural networks have been widely utilized both in adaptive control and for model approximations (Huang et al, 2000(Huang et al, , 2002, thus constituting a worthy area of exploration for force estimation in RAMIS. Li and Hannaford (2017) used a supervised learning technique, Gaussian Process Regression (GPR), to estimate tool-tissue interaction. However, the GPR technique cannot predict well when the target is out of range of the training data.…”