I have a question regarding the ML estimation in presence of colored noise, i.e., Additive Colored Gaussian Noise. I want to implement this using a Viterbi algorithm, since I need to take the 'Correlation Matrix' into account. (Similar to https://dsp.stackexchange.com/a/50777/57695)

The received signal is

$ \mathbf{y} = \mathbf{Hx} + \mathbf{Qn} $,

where $\mathbf{H}$ and $\mathbf{Q}$ are the convolutional matrices of the model, $\mathbf{x}$ is the transmitted signal, and $\mathbf{n}$ is our AWGN. Therefore, $\mathbf{Qn}$ is our ACGN.

How can I generalize the Viterbi algorithm in order to include the correlation matrix of the colored noise in the algorithm?

Thanks a lot

  • $\begingroup$ @Royi Since I couldn't message you, can you please take a look at this? Thank you $\endgroup$
    – Jacob
    Jun 2, 2021 at 2:20


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.