STAG — Sensor-based Tracking and Analysis of Gait

An unsupervised machine-learning pipeline for classifying farmed red deer (Cervus elaphus) behaviour from wearable tri-axial accelerometer data.

STAG discovers prototypical movement patterns directly from sensor streams using k-means clustering, chains them into higher-order behavioural sequences via a Hidden Markov Model, and runs on a 16 MHz microcontroller at over 4 × 10⁸ classifications per second.

Contents

• Installation
  • Requirements
  • With conda (recommended)
  • With pip (CPU only)
  • Verifying the installation
  • Building the documentation
• Usage
  • Stage 1 — Sensor synchronisation
  • Stage 2 — Database construction
  • Stage 3 — GPS trajectory features
  • Stage 4 — GPU clustering
  • Stage 5 — Behavioural analysis
• Pipeline
  • Data collection
  • Feature set
  • Clustering
  • Behavioural prototypes
  • Hidden Markov Model
  • Circadian analysis
  • On-animal deployment
• API Reference
  • stag.sync — Sensor synchronisation
  • stag.database — Database models and ingestion
  • stag.gps — GPS trajectory analysis
  • stag.clustering — k-means clustering and evaluation
  • stag.analysis — Behavioural sequence analysis
  • stag.utils — Utility functions
• Contributing
  • Development setup
  • Code style
  • Running tests
  • Pull requests
  • Reporting issues

Indices and tables

• Index

• Module Index

• Search Page