Force-Sensorless Identification and Classification of Tissue Biomechanical Parameters for Robot-Assisted Palpation.

Area of Science:

• Robotics
• Biomedical Engineering
• Soft Tissue Mechanics

Background:

• Robotic systems are increasingly used in minimally invasive surgery.
• Accurate identification of soft tissue mechanical properties, like stiffness, is crucial for procedures.
• Current methods often require force or tactile sensors and velocity measurements.

Purpose of the Study:

• To develop a method for estimating biomechanical characteristics of soft tissues without force/tactile or velocity sensors.
• To enable model-based state observers using only robot joint positions and commanded torques.
• To enhance diagnostic capabilities for physicians during surgical procedures.

Main Methods:

• A model-based state observer was designed using robot joint positions and commanded torques as inputs.
• Estimated forces and velocities were used for closed-loop force control and mechanical parameter estimation.
• A Bayesian classifier utilized estimated biomechanical parameters for tissue classification.

Main Results:

• A closed-loop force control strategy was implemented using estimated contact forces to prevent tissue damage.
• A least squares estimator was employed to determine mechanical parameters from estimated forces and velocities.
• Combining linear and nonlinear viscoelastic models resulted in 0% misclassification, significantly outperforming linear (50%) or nonlinear (3.12%) models alone for similar samples.

Conclusions:

• The proposed method accurately estimates soft tissue biomechanical properties without traditional sensors.
• The integration of estimated parameters into a Bayesian classifier enhances diagnostic accuracy.
• This approach offers a promising tool for improving safety and efficacy in robotic minimally invasive surgery.