Development and Analysis of Back EMF During User Interaction Dataset for Wired In-Ear Earphones
收藏资源简介：
This dataset contains time-domain feature extractions derived from back electromotive force (back-EMF) signals generated by wired in-ear earphones during three user interaction activities: putting on, taking off, and scratching. The signals were recorded using a non-invasive analog signal acquisition circuit without the use of additional external sensors. Feature extraction was performed directly in the time domain, making the dataset suitable for lightweight signal processing and machine learning applications. Dataset Structure File format: xlsx Number of samples: 350 samples per interaction Number of classes: 3 Class labels: putting_on, taking_off, scratching Features per sample: 24 time-domain features Extracted Time-Domain Features Mean Standard Deviation First Quartile (Q1) Second Quartile / Median (Q2) Third Quartile (Q3) Maximum Value Minimum Value Amplitude Range Number of Prominent Peaks Number of Prominent Valleys Zero Crossings Zero Crossing Rate Maximum Distance Between Prominent Peaks Minimum Distance Between Prominent Valleys Skewness Kurtosis Mean Prominent Peak Value Mean Prominent Valley Value Mean Peak Width Mean Valley Width Root Mean Square (RMS) Variance Signal Energy Signal Entropy Intended Use This dataset can be used for: Exploratory analysis of passive audio-based signals Benchmarking classification algorithms Developing machine learning models for human-device interaction based on earphone back-EMF signals Notes on Data Availability The dataset published on Zenodo represents a curated subset of the full dataset, selected to ensure balanced representation across activities and manageable file size. For access to the complete dataset, please contact the corresponding author at: andriana.kusuma.ft@um.ac.id Related Publication This dataset is associated with the following publication:Development and Analysis of Back EMF During User Interaction Dataset for Wired In-Ear Earphones(https://doi.org/10.1109/ICEEIE66203.2025.11251752)
