IoT-AED — Edge Acoustic Event Detection Dataset
Catalog Summary
"6,400 mono audio clips in seven event classes for edge acoustic-event research, with a three-field manifest and documented reproducibility limitations."
Catalog Notes
Overview
IoT-AED packages environmental audio for acoustic-event detection on IoT edge devices. The publisher describes material derived from UrbanSound8K and additional field recordings.
Files and measured structure
The inspected version 1 archive contains 6,400 WAV files. All headers specify mono, 16-bit PCM at 16 kHz; durations range from 0.05 to approximately 4.037 seconds. The manifest contains filepath, classID and className. Its class counts are horn (claxon): 800; children (niños): 1,000; dog (perro): 1,000; ball (balon): 1,000; siren (sirena): 1,000; ambient (ambiente): 1,000; and people (personas): 600.
Limitations
The distributed manifest has no source-clip, device or session grouping fields, so leakage-safe splitting is not established by the supplied manifest. Some documentation references absent helper/split files and a placeholder Zenodo DOI. The repository declares CC BY 4.0, but dataset_description.txt says “Academic research use.” Source-specific permissions and this inconsistency need clarification before redistribution.
Source and provenance
Dataset DOI: https://doi.org/10.17632/dg5k9pwv4d.1.
View Data Structure
To explore column names, data types, and sample rows, visit the official dataset page on Mendeley Data.
Preview on Mendeley DataCite This Dataset
Quirumbay Yagual, D., & Benalcazar, L. (2026). IoT-AED (Version 1) [Data set]. Mendeley Data. https://doi.org/10.17632/dg5k9pwv4d.1
Source metadata: Mendeley Data (2026) · DOI: 10.17632/dg5k9pwv4d.1
Indexed by IoTDataset.com on Sep 14, 2026
Review the Source Record
Confirm the licence, version, access conditions, file format, and provenance at the source before use.