| Robust Control Toolbox | |
State space realization for a given magnitude Bode plot.
Syntax
[ad,bd,cd,dd,logdfit] = fitd(logd,w) [ad,bd,cd,dd,logdfit] = fitd(logd,w,n) [ad,bd,cd,dd,logdfit] = fitd(logd,w,n,blksz) [ad,bd,cd,dd,logdfit] = fitd(logd,w,n,blksz,flag) [ssd,logdfit] = fitd()
Description
fitd produces a continuous stable minimum-phase state-space realization ssd of a diagonal transfer function matrix such that the diagonal elements' magnitude Bode plots approximately fit Bode magnitude plot data given in the rows of the matrix logd.
Input variable logd is a matrix whose rows are logarithmic magnitude Bode plots evaluated at frequency vector w.
n -- vector containing orders of the state-space approximants of the diagonal scalings (default = 0)
blksz -- a vector of the size of the diagonal blocks (default = 1 for each block). flag -- set to "1" to display a Bode plot (default = 1).
fitd uses the stable routine yulewalk in the Signal Processing Toolbox to fit the continuous magnitude Bode plot. Bilin and polyfit are also used to pre-process the frequency axis shift from s-domain to z-domain. The final discrete realization is transformed back to s-domain via bilin.
fitd plays a crucial role in µ-synthesis design technique. The success of
"D - F" iterations in µ-synthesis depends on the result of fitd.
See Also
bilin, fitgain, invfreqs, musyn, polyfit, yulewalk
| dsigma, sigma | fitgain | |