v0.1.0 — LMS localization milestone

V0.1 milestone

This release completes the clean-room reconstruction's first focused MATLAB milestone:

• configurable 2D microphone geometry and deterministic synthetic signals
• direct-path integer and fractional-delay simulation with optional measured-SNR noise
• independently implemented LMS adaptive filtering
• signed TDOA estimation from learned impulse-response peaks and unwrapped phase slopes
• bounded source-coordinate estimation with visible solver diagnostics
• clean and deterministic noisy end-to-end tests
• executable basic example and reproducible selected figures

Verified result

The selected clean four-microphone example reports an actual source of [0.150000, 0.220000] m, estimated source of [0.146064, 0.214196] m, and localization error of 0.007013 m.

Validation

• MATLAB R2026a Update 4
• 98 passed, 0 failed, 0 incomplete
• Code Analyzer: 0 issues across 45 MATLAB files
• GitHub MATLAB CI passed
• Basic example executed from a clean MATLAB process

Scope

This remains a single-source, synthetic, synchronized, direct-path, two-dimensional signal-processing simulation. It is not AI, a real-time system, a complete room-acoustics model, or a production localization product.

See README.md, docs/METHODOLOGY.md, docs/LIMITATIONS.md, docs/PROVENANCE.md, and docs/V0.1_RESULTS.md.