Reproducible Replication of the Canonical Bicycle Self-Stability (Meijaard 2007)

This release contains a reproducible, hash-verified computational replication of the canonical Whipple bicycle model as formulated in Meijaard et al. (2007), Proc. R. Soc. A. The work independently reconstructs the 2×2 canonical matrices (M, C₁, K₀, K₂), forms the 4×4 state-space system, and computes eigenvalues across a speed sweep using exact-rational SymPy matrices and high-precision mpmath eigensolvers (50-digit evaluation). The classical riderless bicycle self-stability window—where all four eigenvalue real parts are negative—is confirmed at v ∈ [3.65, 5.30] m/s. Resulting eigenvalue tables, stability window files, and checksums are included. A verification script validates both SHA-256 file integrity and agreement with benchmark eigenvalues reported in Meijaard et al. (2007) at v = 4.6 m/s. The dataset is fully self-contained and is intended to serve both as a reference replication of the canonical bicycle model and as the foundation for subsequent TORUS Project investigations into recursive stability frameworks.