Least Squares Fit

In the Least Squares Fit (LSQ) model, the PhiDP unfolding procedure is combined with the conditioning/filtering of PhiDP.

The measured PhiDP are stored as 32-bit unsigned integers. The PhiDP ambiguity interval is mapped linearly into the full range of unsigned integers.

At each bin, an LSQ, weighted with FIR filter coefficients, is performed to compute "conditioned" PhiDP. This LSQ is performed 3 times at each bin, each having a unique phase shift being tracked. One phase shift, φ, is constantly adjusted to keep the current PhiDP in the middle of the ambiguity interval. For the other 2 LSQs, the phases are shifted by the ambiguity interval, φ ± ⅔. The LSQ with the lowest signal-to-noise ratio is assumed to be the unfolded quantity.

The following figure shows the input data and outcomes of this LSQ technique.