Comparing Data Processing and Transmission Scenarios for Spacecraft

Continually increasing amounts of data being generated on spacecraft are driving the need for faster and more power-efficient processing. However, it remains a challenge to determine whether the best option for a given mission scenario is to (1) transmit raw data to ground stations for terrestrial processing, (2) compress data on-board and transmit compressed data to ground stations for terrestrial decompression and processing, or (3) process data on-board and transmit processed data to ground stations. In this study, results are generated and analyzed for timing and energy usage for various mission scenarios using commercial GPUs, military-grade radio transmitters, a neural network application, and both lossy and lossless image compression techniques, all operating on satellite imagery. Results are used to demonstrate how to identify the most efficient data processing options to enable better design and planning for future space missions.