A self-contained automated anomaly detection system for unmanned systems is presented and evaluated. The system and software design provides a platform-agnostic systems health monitoring solution in a small embedded sensing and processing package. Algorithms capable of learning normal operating conditions and detecting anomalies are coupled with a low size, weight, and power hardware implementation to provide a non-invasive, standalone health monitoring capability. Acoustic and inertial sensors (microphone, accelerometer and gyros) were selected as initial non-invasive sensors for detecting a range of common and costly mechanical faults. Test cases are shown to illustrate and validate the system usage and performance.
Number of Pages
Looking for more?
Some of the OnePetro partner societies have developed subject- specific wikis that may help.