PRC-LMS Full Convergence Study: MATLAB Reproducibility Data for Phase-Inversion-Adaptive Vibration Control
Description
This dataset provides MATLAB code, numerical results, reference figures, and Origin-ready plotting data for evaluating the convergence behavior of the parity-resolving cascaded least-mean-square (PRC-LMS) algorithm reported in the associated adaptive-vibration-control article. The reader-guided Live Script first verifies the conventional LMS/NLMS baseline and illustrates the secondary-path phase-inversion behavior. It then presents the complete PRC-LMS sample recursion, representative convergence below and above 90 degrees of lag, internal cascade signals, and primary-filter boundedness. The full convergence envelope is assessed at 50 Hz using independently initialized 100-s cases. Figures 7 and 9 use a 1:5:360 degree diagnostic grid (72 cases), while Figure 8 is independently recomputed at 1:1:360 degree resolution (360 cases). A continuous test varies frequency from 25 to 50 Hz and prescribed lag from 20 to 120 degrees while retaining one adaptive state. The reference non-normalized PRC-LMS learning rates are mu_p = 1e-4, mu_minus = 1e-5, and mu_plus = 2e-5. The package includes a compact MAT archive, tabular summaries, ten expected figures, reusable helper functions, and a self-contained exporter that creates tab-delimited data for every figure for import into Origin. The reported metrics correspond to the deterministic, noise-free, unit-gain fractional-delay implementation domain documented in the package. Very deep fixed-tone attenuation values reflect numerical cancellation approaching floating-point precision and are interpreted separately from experimental noise-floor performance. "Solution to phase inversion problem in adaptive vibration control using parity-resolving cascaded least-mean-square algorithm" Syed Shazaib Shah et al. The article DOI and Mendeley Data DOI will be made available when published.
Files
Steps to reproduce
Read readme file.