arXiv:2611.14225v1 Announce Type: new Abstract: Coherent distributed radar processing requires accurate wireless synchronization. Two-way time transfer (TWTT) obtains reception timestamps from sampled matched-filter peaks, but interpolation bias limits accuracy even when noise is reduced by higher signal-to-noise ratio (SNR) or averaging. Three-point parabolic interpolation is attractive for its simple, explicit formula; reducing its bias generally requires more elaborate fitting, such as higher-order polynomials, or a model tailored to a particular peak shape. A practical alternative should retain this simplicity and interpretability while approaching the low bias of more flexible fits. We propose the Dual-Parabola Interpolator (DPI), which combines two parabolic estimates from five amplitude samples near the matched-filter peak. An adaptive weight cancels the leading interpolation error without requiring a waveform-specific peak model or a correction table. Simulations over flat and standardized multipath channels show bias comparable to a five-point quartic fit, with essentially the same standard deviation as the other interpolators and an execution cost close to a three-point parabolic fit. This combination is useful for TWTT in urban or other multipath settings, where channel distortion prevents reliance on a nominal peak shape.