Edge SpAIce: Enabling Onboard Data Compression With Machine Learning On FPGAs

In the past decade, AI driven algorithms have tremendously increased the reliability in object detection cases for both standard computer vision problems and earth observation image products. Meanwhile, use cases have started to emerge for which a downlink bandwidth limitation should be avoided, for example Space Situation Awareness. This has generated a need for on board AI processing. Earth observation use cases requiring fast responses are also emerging and pushing for reliable algorithms on board, that is to say Deep Neural Networks (e.g. flood detection or early wildfire detection). Consequently, there is a growing need for efficient data processing and compression. Tailoring onboard processing with Machine Learning to specific mission tasks can optimise downlink usage by focusing only on relevant data, ultimately reducing the required bandwidth. The Edge SpAIce project showcases onboard data filtering and reduction by using Deep Learning to identify plastic litter in the oceans. The deployment pipeline, including drastic model compression and deployment using the open-source hls4ml and QONNX tools, enables high-performance, low-power, low-cost computation on onboard FPGA processors. We present lab-based demonstration results, highlighting performance in terms of accuracy, throughput, and power consumption, and discuss planned deployment aspects.