Overview

The Federated Autonomous MEasurement (FAME) project will establish a software framework for connecting dozens spacecraft operated by a diverse cast of private companies and space agencies within a single sensor web. By focusing collections of disparate sensors on a common target, FAME equips researchers with a data set more robust than any data set a single spacecraft could provide. FAME uses decentralized scheduling to automatically allocate observation requests across different commercial and government assets and leverages onboard edge computing to analyze imagery in real-time. This technology streamlines Earth observation, demonstrating the efficacy of autonomous science measurements.

Science Area

Fast-moving events like volcanic eruptions, landslides, flash floods, and harmful algal blooms are difficult to track with traditional space-based sensors. This project will give individual sensors the decision making capabilities they need to automatically focus on rare, short-lived events—such as deep convective ice storms or rapid wildfire expansion. This “dynamic targeting” capability will also allow individual sensors to automatically communicate with one another, coordinating impromptu sensor webs to ensure critical, high-resolution measurements are captured before a time sensitive science event dissipates. Dynamic targeting will also allow sensors to retarget around cloud cover, a common obstacle to obtaining clear Earth observation data.

Technology

In-situ edge computing is key to FAME’s success. FAME leverages Convolutional Neural Networks (CNNs), auto-encoders, and transfer learning to analyze imagery automatically in real time. These onboard ML models perform cloud segmentation, spectral un-mixing, and outlier detection (i.e. identifying active wildfires) without transmitting massive raw data sets to ground stations for human processing. They also allow a sensor within FAME’s sensor web to automatically coordinate with other sensors to collect data on outliers and other time-sensitive features that could be of interest to researchers. FAME’s AI-driven architecture autonomously resolves operational conflicts and triggers rapid follow-up observations, eliminating the need for continuous human intervention.

Advancements

• Federated scheduling architecture translates complex science workflows into automated observation requests, allocating tasks without a central mission controller.
• Onboard Edge Computing utilizes state-of-the-art onboard processors and machine learning models, allowing spacecraft to process data in real time.
• Dynamic targeting allows a spacecraft to automatically adjust its trajectory, reducing interference from cloud cover and enabling rapid follow-up observations.

Principal Investigator

Dr. Steve Chien is JPL Fellow, Senior Research Scientist, and Technical Group Supervisor of the Artificial Intelligence Group and in the Mission Planning and Execution Section at the Jet Propulsion Laboratory, California Institute of Technology where he leads efforts in automated planning and scheduling for space exploration. Dr. Chien was previously Adjunct Faculty with the Department of Computer Science of the University of Southern California, and a Research Scientist at the Joint Institute for Regional Earth System Science & Engineering and a Visiting Scholar with the Department of Computer Science of the University of California at Los Angeles. He holds a B.S. with Highest Honors in Computer Science, with minors in Mathematics and Economics, M.S., and Ph.D. degrees in Computer Science, all from the University of Illinois.

Selected Publications

1. Zilberstein, I.; and Chien, S. Large-Scale Continual Scheduling and Execution for Dynamic Distributed Satellite Constellation Observation Allocation (Extended Abstract). In Proc. of the 25th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2026), May 2026.
2. Chien, S.; Zilberstein, I.; Candela, A.; Barretta, D.; Rijlaarsdam, D.; Hendrix, T.; Dunne, A.; Grau, O. C.; i Mestre, A. G.;  Bovez, M. P.; Aragon, O.; Miquel, J. P.; Mogannam, J.; Scher, M.; van Duijn, P.; Subramanian, A.; Vatsal, V.; and Kothandhapani, A. New Observing Systems Flight Demonstration (Extended Abstract). In Proceedings of 18th Symposium on Advanced Space Technologies in Robotics and Automation, October 2025.
3. Zilberstein, I.; Candela, A.; and Chien, S. Federated Autonomous Operations: A New Paradigm for Large-Scale Observation Systems. In Proceedings of the 18th International Conference on Space Operations, May 2025.
4. Chien, S.; Zilberstein, I.; Candela, A.; Barretta, D.; Rijlaarsdam, D.; Hendrix, T.; Dunne, A.; Grauc, O. C.; i Mestrec, A. G.; Bovec, M. P.; Aragon, O.; Miquel, J. P.; Subramanian, A.; Vatsal, V.; Kothandhapani, A.; Mogannam, J.; and Scher, M. Multi-Asset New Observing Systems Flight Demonstration. In Proceedings of the 18th International Conference on Space Operations, May 2025.
5. Zilberstein, I.; Rao, A.; Salis, M.; and Chien, S. Decentralized, Decomposition-Based Observation Scheduling for a Large-Scale Satellite Constellation. Journal of Artificial Intelligence Research, 82: 169–208. January 2025.