For the state-space model with colored measurement noise, the traditional measurement difference method can transform the colored measurement into a white measurement with a one-step delay, but when the colored noise has a non-zero mean value, the mean value of this method changes after propagation. As a result, the accuracy of the filter is reduced. How can we model to solve this problem?
Dear Zihao Jiang
The random process X(t) is called a white Gaussian noise process if X(t) is a stationary Gaussian random process with zero mean, μX=0, and flat power spectral density, SX(f)=N02, for all f. This again confirms that white noise has infinite power, E[X(t)2]=RX(0).
Kind Regards
Qamar Ul Islam
You do not indulge what your signal source is - photos? (ie image) or something more exotic...?
- Badges
- Science topic
- Similar topics
- Control Systems Engineering
- Filtering