Table 7.

Typical parameter identification methods.

Item	Typical Method		Advantages and Limitations		Applications
					Needle Deflection & Tissue Deformation	Path Planning & Navigation	Force Analysis	Online Force Control
Data-based parameter identification	System response method; Frequency response method; Correlation method; Maximum likelihood method		Acquire data distribution characters; Reflect specific criteria; Data correlation analysis	Require high integrity; Huge workload; Offline estimation	√	√	√	×
	Static system	Dynamic system
Time-invariant parameter identification	Weighted least-squares estimation; Constrained least-squares estimation; Truncated least-squares estimation; Total least-squares estimation; Nonlinear least-squares estimation	Least-squares estimation; Ordinary least-squares estimation; Biased least-squares estimation; Generalized least-squares method; Pre-filtering method; Neural network; Wavelet network	Characterize the entire system simply	Not well reflect the real situation of the whole system	√	×	√	×
Time-varying parameter identification	Recursive least-squares estimation; Square root filtering; Reduced-rank square root (RRSQRT) filtering; Extended Kalman filtering for the estimation	Recursive prediction-error estimation; Fixed-interval optimal smoothing; Extended Kalman filtering; Neural network; Wavelet network; Radial basis function neural network; Genetic algorithm; Evolutionary algorithm; Fuzzy logic algorithm; Times series analysis method	Online control; reflect the dynamic characteristics	Improve the complexity of analysis and research	√	√	√	√

In the column of Applications, “√” means yes, “×” means no.