Font Size:
Bayesian approach to bilinear system identification
Last modified: 2017-05-25
Abstract
The paper discusses the Bayesian parameter estimation of a piecewise linear system. This involves estimating the unknown
properties of the stiffness characteristics as well as the distance at which the change in stiffness characteristic takes place. This
problem is motivated by a medical application in which a palpation device is developed in an effort to substitute a manual
palpation procedure. The applied methodology relies on a recently developed approach by Dr. Green, which allows the tracking
of parameter estimates as the information from an increasingly large data set is analysed. The paper highlights the importance of
establishing when a sufficient amount of training data has been utilised to allow convergence of parameter estimates
properties of the stiffness characteristics as well as the distance at which the change in stiffness characteristic takes place. This
problem is motivated by a medical application in which a palpation device is developed in an effort to substitute a manual
palpation procedure. The applied methodology relies on a recently developed approach by Dr. Green, which allows the tracking
of parameter estimates as the information from an increasingly large data set is analysed. The paper highlights the importance of
establishing when a sufficient amount of training data has been utilised to allow convergence of parameter estimates