Footstep-Induced Floor Vibration Data for Person Identification with 8 Walking Speeds

Stanford University researchers collected structural vibration data from 10 participants walking in a 30ft x 6ft concrete hallway. The dataset contains footstep traces recorded by 5 geophone sensors at a 1000Hz sampling rate, with labels for person identity, sensor number, and 8 distinct walking speeds. It was published in 2021 and is stored in a MAT-file format.

Use Cases

• Training person identification models based on unique footstep-induced vibration patterns.
• Studying the impact of walking speed variation on structural vibration signatures.
• Developing sensor fusion techniques using data from multiple geophone sensors.
• Benchmarking time-series classification algorithms on labeled, multi-factor sensor data.

Strengths

• Data from 10 distinct participants provides a basis for multi-class identification tasks.
• Includes 8 systematically varied walking speeds per person, allowing for robustness testing.
• Recorded by 5 synchronized geophone sensors, enabling multi-perspective analysis.
• Sampling frequency of 1000Hz captures detailed temporal vibration signals.

Limitations

• Row count and dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data collection involved a specific 30ft x 6ft concrete hallway, which may limit generalizability to other environments.

Provenance

Source

Stanford University researchers, published via Zenodo.

Collection Method

Experimental data collected from 10 participants walking in a controlled hallway environment, recorded by geophone sensors.

Geography

Data was collected in a controlled indoor hallway environment; specific geographic location is not stated.