Dynamic modeling of breast tissue with application of model reference adaptive system identification technique based on clinical robot-assisted palpation

• Fabrication of a robot-assisted device to accurately mimic physician׳s palpation.
• Considering a time-dependent viscoelastic behavior for the breast tissue.
• Performing breast clinical examination by the device based on Helsinki agreement.
• Estimation of a dynamic model representative of tissue׳s behavior based on MRAS.
• Measurement of error signal between dynamic model output and experimental dataset.

Accurate identification of breast tissue׳s dynamic behavior in physical examination is critical to successful diagnosis and treatment. In this study a model reference adaptive system identification (MRAS) algorithm is utilized to estimate the dynamic behavior of breast tissue from mechanical stress–strain datasets. A robot-assisted device (Robo-Tac-BMI) is going to mimic physical palpation on a 45 year old woman having a benign mass in the left breast. Stress–strain datasets will be collected over 14 regions of both breasts in a specific period of time. Then, a 2nd order linear model is adapted to the experimental datasets. It was confirmed that a unique dynamic model with maximum error about 0.89% is descriptive of the breast tissue behavior meanwhile mass detection may be achieved by 56.1% difference from the normal tissue.