idpoly: Polynomial model with identifiable parameters

In sysid: System Identification in R

Description

Creates a polynomial model with identifiable coefficients

Usage

idpoly(A = 1, B = 1, C = 1, D = 1, F1 = 1, ioDelay = 0, Ts = 1,
  noiseVar = 1, intNoise = F, unit = c("seconds", "minutes", "hours",
  "days")[1])

Arguments

A	autoregressive coefficients
B, F1	coefficients of the numerator and denominator respectively of the deterministic model between the input and output
C, D	coefficients of the numerator and denominator respectively of the stochastic model
ioDelay	the delay in the input-output channel
Ts	sampling interval
noiseVar	variance of the white noise source (Default=1)
intNoise	Logical variable indicating presence or absence of integrator in the noise channel (Default=FALSE)
unit	time unit (Default="seconds")

Details

Discrete-time polynomials are of the form

A(q^{-1}) y[k] = \frac{B(q^{-1})}{F1(q^{-1})} u[k] + \frac{C(q^{-1})}{D(q^{-1})} e[k]

Examples

# define output-error model
mod_oe <- idpoly(B=c(0.6,-0.2),F1=c(1,-0.5),ioDelay = 2,Ts=0.1,
noiseVar = 0.1)

# define box-jenkins model with unit variance
B <- c(0.6,-0.2)
C <- c(1,-0.3)
D <- c(1,1.5,0.7)
F1 <- c(1,-0.5)
mod_bj <- idpoly(1,B,C,D,F1,ioDelay=1)

sysid documentation built on May 2, 2019, 4:18 a.m.