Anomaly detection on ESP32 with 24 bytes of RAM - no cloud, no malloc

This article provides a detailed overview of Ardent Pulse, an anomaly detection SDK designed for microcontrollers (MCUs) with extremely low resource requirements. Here are key points and insights from the content:

1. Purpose:

   • The SDK aims to detect anomalies in sensor data on MCUs without requiring internet connectivity or cloud services.

2. Resource Constraints:

   • Zero heap allocations
   • Less than 4 KB RAM per detector
   • Detection time of less than 1 ms at 80 MHz clock speed

3. Technologies Used:

   • C99 standard for portability across different MCUs and environments.
   • No dependency on external libraries or frameworks, ensuring minimal resource usage.

4. Key Components:

   • Z-Score Detector: Measures how many standard deviations away from the mean a data point is.
   • EWMA (Exponentially Weighted Moving Average) Drift Detector: Tracks gradual changes in sensor readings over time.
   • MAD (Median Absolute Deviation): Robust to outliers, suitable for non-Gaussian distributions.

5. Implementation Details:

   • Each detector is implemented as a small struct with minimal memory footprint.
   • The code is

Read the full article at DEV Community

---

Want to create content about this topic? Use Nemati AI tools to generate articles, social posts, and more.