What would you do with access to high-performance computing power in space? Better yet, how could you use that power to benefit life on Earth, improve monitoring of our planet’s ever-changing environment, and even advance our capabilities for space exploration?
We were given that very opportunity, and based on our expertise in remote sensing, our answer was to deploy a full enterprise version of the industry leading image processing and analysis software ENVI in space. This recently launched ENVI is Space mission is operated by HySpeed Computing in partnership with L3Harris Geospatial and Hewlett Packard Enterprise (HPE). With an initial focus on demonstrating an array of applications in Earth observation, ENVI in Space will also showcase capabilities in other fields, such as monitoring plant health for food cultivation on long-duration space missions and analyzing the characteristics and composition of Lunar and Martian surfaces. The possibilities are endless.
To achieve these objectives, ENVI has now successfully been installed and tested on the HPE Spaceborne Computer-2 (SBC-2), which is the first commercial off-the-shelf edge computing system with AI capabilities built for in-space data processing. The system was installed and powered up on the International Space Station (ISS) in May 2021. HPE Spaceborne Computer-2 builds on the success of the first Spaceborne Computer, which HPE designed and launched to space from 2017-2019 as a proof of concept to test if commercial, off-the-shelf computers can withstand the impact of a rocket launch and operate in the harsh conditions of space, such as unpredicted levels of radiation and lack of gravity.
While onboard space-based computing is not a new concept, never before has such a powerful computing system been deployed in space. As the field of space computing continues to expand its horizons, SBC-2 and ENVI in Space offer worlds of opportunity for exploring new capabilities in edge computing.
What are some of the advantages that such technology provides? First, there’s the efficiency and speed of data analysis. By moving the compute resources and software algorithms closer to the data, analysis can be performed at the source and results downlinked in real-time as completed products. This has clear benefits for time critical applications such as disaster preparedness and response.
These same real-time results can also be utilized directly for other onboard applications, such as sensor operation and image acquisition. For example, imagery can be analyzed as it’s acquired, and repeat acquisitions can be iteratively scheduled until certain conditions are met, such as a cloud cover threshold is met, or a pre-defined amount of change is observed in a given area of interest. Or, when an event occurs, results from one sensor could be used to direct acquisitions of other sensors with different resolutions or modalities to automatically obtain additional information characterizing the event.
Second, there’s the flexibility and breadth of capabilities. Rather than deploying software customized for specific applications, the combination of enterprise software and computing provides a full suite of image processing and analysis capabilities. This allows applications to rapidly adapt to new situations or needs, as well as rapidly deploy updates or new algorithms to an existing system. This becomes particularly important in remote locations, such as space exploration or missions to the Moon and Mars, where image analysis demands are both more diverse and more difficult to anticipate. For example, inspecting exposed surfaces for damage or contamination, monitoring infrastructure, or performing medical diagnosis.
Third, there’s technology demonstration and risk reduction. Despite the profound advances that continue to improve our access to space, working in space is still a challenge. Validating system performance and assessing software execution in an operational environment can significantly reduce risk before deploying payloads on independent free-flying satellites. The enterprise solution provided by ENVI in Space thereby provides a robust testbed for evaluating any number of other payloads and applications.
And fourth, but certainly not the least or last advantage, there’s the functionality of acting as a centralized hive-mind or hive-server for distributed sensing systems. Consider, for example, a system of cubesats, drones or rovers, each with their own sensors and acquisition assignments. Not only can a centralized computer provide operations management for this system, but it can also serve as the backbone for storing, processing, analyzing, and distributing the collected data.
We’re sure that you also have your own ideas of what you’d do with this technology. Share your thoughts below and just maybe we’ll have the chance to make these ideas a reality.
Additionally, in the coming months we’ll continue putting the ENVI in Space mission through its paces, so stay tuned for future updates.
images: NASA (top), HySpeed Computing (body)
