Overview
Satellites collect various amounts of telemetry data that is then downlinked to Earth and processed. This enables closer monitoring of satellite status and ensures that the correct actions are taken to keep the satellite operational as long as possible. Anomaly detection is crucial in this process, as timely warnings to the operations team ensure that necessary corrections are executed promptly. On-board satellite processing is limited due to resource, power, and downlink constraints. Most satellites use some form of FPGA for such tasks, and we deployed our custom algorithms for this purpose.
Goals
The goal was to port our existing anomaly detection solution to an FPGA environment similar to that of OPS-SAT. The solution needed to adhere to specific requirements to interact correctly with the satellite platform.
Solution
We ported our existing anomaly detection solution to the FPGA environment, developing our own IP cores that utilized the parallel processing capabilities of FPGAs while managing the resource constraints of the specific FPGA. This required significant optimization and streamlining of the processing. In addition to custom IP cores, we mapped the device tree of our FPGA module to successfully access it from the on-board Linux operating system. A complementary C code was developed to integrate the entire system, ensuring robustness and reliability.
Results
- Successful deployment on-board the satellite
- The algorithm was executed successfully
- The detection results were successfully downlinked
