Model description:
$$\begin{align*} x_1(t+1) &=\left(\dfrac{x_1(t)}{1+x_1^2(t)}+1\right)\sin{x_2(t)} \\ x_2(t+1) &=x_2(t)\cos{x_2(t)}+x_1(t)e^{-((x_1^2(t)+x_2^2(t))/8} + \dfrac{u^3(t)}{1+u^2(t)+0.5\cos{x_1(t)+x_2(t)}} \\ y(t) &=\dfrac{x_1(t)}{1+0.5\sin{x_2(t)}}+\dfrac{x_2(t)}{1+0.5\sin{x_1(t)}}+e(t), \end{align*}$$
where $e(t)$ is the noise term, has a variance of 0.1.
Type:
Form:
Model order:
2
Time domain:
Linearity:
Publication details:
| Title | Nonlinear system identification via direct weight optimization |
| Publication Type | Journal Article |
| Year of Publication | 2005 |
| Authors | Roll, Jacob, Nazin Alexander, and Ljung Lennart |
| Journal | Automatica |
| Volume | 41 |
| Pagination | 475 - 490 |
| Date Published | 01/2005 |
| ISSN | 0005-1098 |
| URL | http://dx.doi.org/10.1016/j.automatica.2004.11.010 |
