This repository provides a geometric alignment analysis of VLBI (Very Long Baseline Interferometry) NGS observation data. Unlike standard post-processing methods that rely on extensive parameter tuning and manual residual shaving, this project focuses on verifying the fundamental geometric consistency of space-time using minimal variables.
The analysis adopts a minimalist approach to preserve the integrity of raw observation data. Instead of utilizing hundreds of atmospheric and instrumental correction parameters, we apply two core geometric constants:
- ck (Absolute Speed of Light): 297,880,197.6 m/s
- S_earth (Scale Factor): 1.006419562
The engine performs a direct geometric strike without any post-hoc manual adjustments. Our goal is to observe the raw convergence of data when aligned to a deterministic geometric frame.
- No Manual Shaving: We do not artificially "shave" residuals to force-fit the data into an idealized curve.
- Geometric Causality: We prioritize the causal relationship between geometry and time-delay over statistical precision achieved through overfitting.
- System U Alignment: The observation stations and sources are mapped into an absolute geometric coordinate system to evaluate the "Raw Truth" of the residuals.
- Dataset: 20JAN02XE_N005.ngs
- Median Residual: ~108,432 ns
- Observation: By applying only ck and S_earth, the chaotic raw data (initially exhibiting ~20ms error) collapses into a deterministic band of approximately 0.1ms. This residual is maintained as a "honest scar"—a reflection of uncorrected local gravitational and instrumental variables—rather than being suppressed through artificial post-processing.
This project demonstrates that a significant portion of what is often considered "stochastic noise" can be resolved through fundamental unit redefinition and geometric alignment. While standard models achieve sub-nanosecond precision through thousands of ad-hoc corrections, this analysis suggests that the underlying geometric structure of the universe is far more consistent than complex post-diction models imply.
"We do not predict the past; we analyze the causality of the future."
Lead Researcher: estake@naver.com