Project:	Project Title: AI for Sentinel-6 Instrument Health Monitoring (JPL AO No. 16881)
Disciplines:	Artificial Intelligence, Computer Science
Mentor:	Umaa Rebbapragada, (JPL), Umaa.D.Rebbapragada@jpl.nasa.gov, Phone: (818) 354-0038
Background:	JPL is actively pursuing AI/ML solutions to enhance efficiency and mitigate risks in mission operations. A readily achievable application involves using anomaly detection and supervised learning for the routine monitoring of spacecraft telemetry. Given that hundreds of telemetry channels are downlinked, manual review for signs of degrading spacecraft or instrument health is impractical and costly. Our project aims to apply AI/ML methods across all channels to flag both known and unknown conditions that may indicate problems for mission operators. Our overall project goal is to deliver a tool that offers clear value to current operations teams that ensures a comprehensive and consistent monitoring of telemetry channels at low false positive and negative rates. Our project will establish the necessary infrastructure to create ML-ready datasets from mission data archives, advance ML methods on targeted use cases, integrate into current operations tools and validate to gain user acceptance. Our long-term goal is to demonstrate multi-mission applicability.
Description:	As the project team works to advance the goals described above, the intern will be performing guided research tasks related to advance the state of the data science methods used. The intern will be helping with the following tasks: 1) establishing the metrics and validation benchmarks used to evaluate the methods used in this project, and 2) developing new methods that can plug into and be evaluated against the benchmark, and 3) marking improvements to the benchmark as we discover that we must enhance the current validation sets in response to feedback from our operations stakeholders. The selection of methods to evaluate will be decided by the intern and the project manager. Some of the research questions we are interested in investigating are the use of open source frameworks for time series anomaly detection and supervised learning as well as the use of time series foundation models. We are also interested in investigating the use of Generative AI frameworks in order to enhance the operator experience with natural language style prompting or curate training data from unstructured text.
References:	Spacecraft Time-Series Anomaly Detection Using Transfer Learning (CVPR Workshop 2021): https://openaccess.thecvf.com/content/CVPR2021W/AI4Space/papers/Baireddy_Spacecraft_Time-Series_Anomaly_Detection_Using_Transfer_Learning_CVPRW_2021_paper.pdf A Review of Anomaly Detection in Spacecraft Telemetry Data: https://www.mdpi.com/2076-3417/15/10/5653 NeuralProphet: Explainable Forecasting at Scale: https://arxiv.org/abs/2111.15397 When Foundation Models are One-Liners: Limitations and Future Directions for Time Series Anomaly Detection: https://openreview.net/pdf?id=H27kvyG4qf
Student Requirements:	Core Coursework: Completion of advanced Machine Learning coursework, including all mathematical foundations (Calculus, Multi-variable Calculus, Linear Algebra, and Probability & Statistics). Computer Science Foundations: Strong understanding of Data Structures and Algorithms. Programming: Proficiency in Python is required, including hands-on experience with popular open-source machine learning frameworks (e.g., PyTorch, TensorFlow, Scikit-learn). Software Engineering: Familiarity with software engineering principles and version control. Active experience using GitHub for collaborative development is expected. Self-Management: Ability to self-organize, pace individual research tasks, and maintain productivity in a hybrid or independent environment. Communication: Strong verbal and written communication skills; responsible for providing concise weekly status updates to project stakeholders. Preferred Qualifications: Advanced coursework in Deep Learning or specialized AI topics. Direct experience with Time Series analysis and anomaly detection methods.
Location / Safety:	Project building and/or room locations: . Student will need special safety training: .
Programs:	This AO can be done under the following programs:   Program    Available To        SURF@JPL    both Caltech and non-Caltech students  Click on a program name for program info and application requirements.