Article

Open-source tool SONAR: an analysis of SGD-based online one-class joint anomaly detection

The Dense-Armor project has open-sourced the SONAR (SGD-based One-Class Novelty detection with Approximate RBFs) module, which uses random Fourier features (RFF) to approximate an RBF kernel and adds strongly convex regularization, allowing the one-class boundary to be updated per sample via online SGD to detect whether robot joint residuals deviate from a healthy state.

Part of the event: Dense-Armor Open-Sources SONAR, an SGD-Based One-Class Joint Anomaly Detection Module

This page provides a summary and source link only.