Algorithm 1: Proposed MF method
Input:	Incoherent SBEs: $M_i(n), M_j(n).$
Step 1:	Estimate the linear phase $\widehat{\gamma }$ between two SBEs and compensate for it to $M_j(n)$, obtain ${\widehat{M}}_j(n)$.
Step 2:	Extract SCs from $M_i(n)$ and ${\widehat{M}}_j(n)$ based on the GTD model, obtain the poles and the complex amplitude: $\{p_{mi},d_{mj}\},\{p_{mj},d_{mj}\}$ and categorize the SCs into two types: ISCs: $\{p_{miISC},d_{miISC}\},\{p_{mjISC},d_{mjISC}\}$ and USCs: $\{p_{miUSC},d_{miUSC}\},$ $\{p_{mjUSC},d_{mjUSC}\}.$
Step 3:	For ISCs, calculate the new fixed phase ${\widehat{\beta }}_{mISC}$ for each ISC and compensate for it to each ISC of SubJ, obtain the new complex amplitude: ${\widehat{d}}_{mjISC}$.
Step 4:	Reconstruct the SBEs of ISCs based on the $\{p_{miISC},d_{miISC}\}$ and $\{p_{mjISC},{\widehat{d}}_{mjISC}\}$ to obtain $M_{iISC}\left(n\right)$, $M_{jISC}\left(n\right)$.
Step 5:	Applying the second step and the third step of MF to generate the UWBE of ISCs.
Step 6:	For USCs, apply the band extrapolation to generate the corresponding UWBEs of USCs for each sub-band.
Step 7:	Add the UWBE of ISCs and UWBE s of USCs to generate the final estimated UWBE.
Output:	The final estimated UWBE.