What is it about?
In this article, we describe RAMAN - a framework of approaches for anomaly detection in telemetry data of Mars rover power subsystems that detect anomalies across pair signals, single signals, battle shorts, sensors, and equipment configuration. Novel approaches like using spline interpolation on the derivative of error calculation, and large-margin cosine loss help us achieve accurate results across various rover configurations.
Featured Image
Photo by Sufyan on Unsplash
Why is it important?
This work can provide you with a variety of solutions to tackle your anomaly detection problems depending on the type of data. It also provides a commentary of insights on the challenges to be mindful of when working in this domain.
Perspectives
This was an interesting problem to tackle given the focus on speed, precision (fewer false positives), and statistical approaches for anomaly detection.
Pratik Ratadiya
University of California San Diego
Read the Original
This page is a summary of: RAMAN: Robust Approaches for Multimodal ANomaly detection in Mars Rover Power Systems, January 2024, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/6.2024-1378.
You can read the full text:
Resources
Contributors
The following have contributed to this page
