.Noise.UsersGuide.QuickStartGuide

Information

To model a noise source in using this library, the u the block Noise.PRNG has to be selected. properly parametrizing this block is done on two levels: easy and advanced:

Easy Noise

Select a (Pseudo-) Sampling time for the block. This limits the bandwith of the block to 1/sampling time. A constant spectral density will be used. This type of noise will be good for standard sensor noise.

Advanced Noise

To select more advanced noise methods with a defined Probability Density Function to define the distribution of the generated random values, the "Advanced" Tab has to be selected. The user can select a wide range of distributions:

• Uniform
• Normal distribution
• Couchy-Lorenz distribution
• Weibull distribution
• Irwin-Hall distributiuon
• Bates distribution - Suitable for mean-filtered noise signals
• Discrete distribution - Suitable for Spike noise

Furthermore the Power Spectral Density can be set to specify the frequency characteristics of the signal. Following methods are supplied:

• Unmodified white noise with constant power spectral density
• low pass filtered noise
• linear interpolation between noise samples
• interpolation using convulution with a Kernel
• filter by convulution with a impulse response

Using the modular approach of the fuctions, it is possible for the user to develop and select their own kernals for specialized frequency responses.

For more information, please also see Noise.PRNG.SampleBasedRNG:

Revisions

Developed 2014 at the DLR Institute of System Dynamics and Control