Abstract:Artificial intelligence (AI) and autonomous edge computing in space are emerging areas of interest to augment capabilities of nanosatellites, where modern sensors generate orders of magnitude more data than can typically be transmitted to mission control. Here, we present the hardware and software design of an onboard AI subsystem hosted on SpIRIT. The system is optimised for on-board computer vision experiments based on visible light and long wave infrared cameras. This paper highlights the key design choices made to maximise the robustness of the system in harsh space conditions, and their motivation relative to key mission requirements, such as limited compute resources, resilience to cosmic radiation, extreme temperature variations, distribution shifts, and very low transmission bandwidths. The payload, called Loris, consists of six visible light cameras, three infrared cameras, a camera control board and a Graphics Processing Unit (GPU) system-on-module. Loris enables the execution of AI models with on-orbit fine-tuning as well as a next-generation image compression algorithm, including progressive coding. This innovative approach not only enhances the data processing capabilities of nanosatellites but also lays the groundwork for broader applications to remote sensing from space.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper attempts to address the challenges of deploying artificial intelligence (AI) systems on nanosatellites. Specifically, the paper focuses on the following aspects: 1. **Robustness in Extreme Space Environments**: - The space environment is extremely harsh, including radiation from high-energy protons and low-energy electrons, as well as extreme temperature variations. These factors pose threats to the satellite's electronic equipment and data integrity. The paper proposes a series of designs and strategies to ensure that the AI system can operate stably in such an environment. 2. **Limited Computational Resources**: - Due to their small size and low cost, nanosatellites have very limited computational resources. This limits the complexity of AI models that can be deployed and their ability to be fine-tuned and retrained in orbit. The paper optimizes hardware and software design to enable the AI system to operate efficiently with limited computational resources. 3. **Limited Communication Bandwidth**: - The communication bandwidth between nanosatellites and ground stations is very limited, usually with only a few short transmission windows each day. The amount of data generated by modern sensors far exceeds the amount that can be transmitted. The paper proposes an efficient image compression algorithm (JPEG-XL) to reduce data transmission volume and uses progressive encoding techniques to achieve partial data transmission, thereby improving bandwidth utilization. 4. **Autonomy and Adaptability**: - AI systems operating in orbit need to have a high degree of autonomy because communication is limited, making frequent interactions with ground stations impossible. The paper designs an adaptive AI system that can maintain robustness when encountering unexpected environmental changes and data variations, and can be fine-tuned and retrained in orbit. ### Summary The paper demonstrates how to successfully deploy and operate complex AI systems on nanosatellites by designing and optimizing a payload named "Loris." The Loris payload integrates an NVIDIA Jetson Nano single-board computer and a multi-camera system, aiming to address the aforementioned challenges, enhance the data processing capabilities and real-time decision-making of nanosatellites, and lay the foundation for future space remote sensing applications.

Mitigating Challenges of the Space Environment for Onboard Artificial Intelligence: Design Overview of the Imaging Payload on SpIRIT

SpIRIT Mission: In-Orbit Results and Technology Demonstrations

Unlocking the Potential of Small Satellites: TheMIS's Active Cooling Technology on the SpIRIT Mission

Improving Performance-Power-Programmability in Space Avionics with Edge Devices: VBN on Myriad2 SoC

Onboard Science Instrument Autonomy for the Detection of Microscopy Biosignatures on the Ocean Worlds Life Surveyor

Artificial Intelligence Satellite Telecommunication Testbed using Commercial Off-The-Shelf Chipsets

VERTECS: A COTS-based payload interface board to enable next generation astronomical imaging payloads

The Space Infrared Interferometric Telescope (SPIRIT): High-resolution imaging and spectroscopy in the far-infrared

Biological research and self-driving labs in deep space supported by artificial intelligence

Characterisation of SATCOM Networks for Rapid Message Delivery: Early In-Orbit Results

Infrared Remote Sensing Using Low Noise Avalanche Photodiode Detector

Autonomous Payload Thermal Control

Mechanical design and development of SPRITE: a 12U CubeSat with a far-ultraviolet imaging spectrograph

Efficient onboard multi-task AI architecture based on self-supervised learning

Challenges of Artificial Intelligence in Space Medicine

Beyond Low Earth Orbit: Biomonitoring, Artificial Intelligence, and Precision Space Health

Reaching the Edge of the Edge: Image Analysis in Space

Beyond Low Earth Orbit: Biological Research, Artificial Intelligence, and Self-Driving Labs

Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning

Artificial Intelligence Frameworks to Detect and Investigate the Pathophysiology of Spaceflight Associated Neuro-Ocular Syndrome (SANS)

Inter-Satellite Communication System based on Visible Light