pygbn

This package implements the generalized binary noise (GBN) model of [1] in Python. The code is based on the Matlab implementation revised by Ivo Houtzager in 2007 at the Delft Center of Systems and Control.

Installation

You can very easily install the package using pip:

pip install pygbn

or after locally cloning the source code:

pip install .

Usage

Below is an example of how to use the package.

import matplotlib.pyplot as plt
from pygbn import gbn

if __name__ == '__main__':
    seed = 0 # random seed

    h = 0.05 # sampling period [s]
    T = 1 # length of signal [s]
    A = 1 # amplitude of signal
    ts = 1 # estimated settling time of the process [s]

    # flag indicating process damping properties
    # flag = 0 if the process is over-damped (default)
    # flag = 1 if the process is oscillary (min phase)
    # flag = 2 if the process is oscillary (non min phase)
    flag = 0

    # generate time array
    t = np.arange(start=0, stop=T, step=h)

    # generate the signal
    # the gbn function returns a time array and a signal array
    u = gbn(h, T, A, ts, flag, seed=seed)

    # optional: plot the generated signal
    plt.plot(t, u)

Citations

[1] Tulleken, H. J. (1990). Generalized binary noise test-signal concept for improved identification-experiment design. Automatica, 26(1), 37-49.