Calculating Optimal Weights for MVDR Signal Processing Gain

[nikki92]

Using a MVDR method for a uniformly spaced linear array of sensors, I calculated the optimal weights. The procesing gain is (SIR)out / (SIR)in . The value turned out complex. Is this possible/makes sense?

 

[meBigGuy]

A phase shift would be represented by a complex gain.

 

[nikki92]

Thanks! So the power is simply represented by the real part?

 

[meBigGuy]

No. The magnitude of the output determines the power. Complex numbers can be thought of as real and imaginary (in-phase and quadrature components), or as a magnitude at a phase angle.

 

[alphy]

Yes, it is possible and makes sense for the processing gain to have a complex value in the case of MVDR signal processing. This is because the MVDR method takes into account the complex spatial characteristics of the signals received by the array of sensors. The optimal weights are calculated based on the complex spatial correlation matrix of the signals, which can result in a complex processing gain. This complex processing gain accounts for both the amplitude and phase differences between the desired signal and the interfering signals, allowing for improved signal quality and noise reduction. Therefore, the complex value of the processing gain is a reflection of the sophisticated and effective nature of the MVDR method in signal processing.