Baby Model Zoo

Small, specialized EdgeAI model projects for resource-constrained embedded systems.

Penguin temperature anomaly detector

Penguin combines a temperature autoencoder and predictor with an NXP FRDM-MCXC162 firmware target, host tests, flash logging, and a browser dashboard. The project samples the board sensor, maintains a 16-reading window, trains against a selected baseline, and reports ready, watch, or anomaly states.

Penguin temperature anomaly dashboard visual

Setup

  1. Read the project README and its firmware setup.
  2. Use the documented Node.js 24 dashboard prerequisites and board-specific toolchain.
  3. Run the documented host tests before attempting hardware validation. Hardware timing, false-alert rate, latency, LED behavior, and power-loss behavior remain project-specific validation items.

Penguin is experimental environmental monitoring, not a medical device or safety alarm. See MRD for hardware facts and EIL for bounded runtime context.