Deployability reporting audit dataset for "Edge-Device Deployable Self-Supervised Traversability Estimation for Field Robots: A Critical Review"

This dataset contains the per-paper coded indicators from the structured deployment-constraint reporting audit reported in the review article "Edge-Device Deployable Self-Supervised Traversability Estimation for Field Robots: A Critical Review." Version 2 adds bibliographic metadata. Each record now carries authors, year, title, venue, and DOI columns, so every audited paper can be traced to the published literature. Version 1 identified papers by source filename only. No coded indicator value was changed: the v2 files are identical to v1 in every original column, and all reported percentages are unaffected. The DOI column is partially populated (61 of 252 rows), covering records independently verified against Crossref; the remainder are identifiable from authors, year, title, and venue. The audit covers the deployability-relevant subset of the surveyed corpus: the union of three thematic populations — segmentation/quantization for edge neural processing units (Theme 2, 81 papers), self-supervised label generation (Theme 3, 73 papers), and multi-modal sensor fusion under embedded compute (Theme 4, 98 papers) — comprising 252 paper-appearances and 228 unique papers after deduplication. Each paper was coded for whether it reports four deployment constraints: power envelope, thermal behaviour under sustained load, quantization-induced accuracy degradation, and edge-hardware inference latency. Files: meta_theme2.csv: per-paper coded indicators, Theme 2 (segmentation/quantization). meta_theme3.csv: per-paper coded indicators, Theme 3 (self-supervised labels). meta_theme4.csv: per-paper coded indicators, Theme 4 (multi-modal fusion). meta_extraction_summary.md: aggregate counts, per-theme breakdowns, and the headline reporting rates, with the audit's coding rules and limitations. README.txt: column definitions, methodology, and citation guidance. Coding was conservative (ambiguous cases scored negative), which biases the reported rates downward. See Sections 2.2 and 8.1 of the article for full methodology.