Session
Swifty Session 6: Ground Systems & Operations
Location
Utah State University, Logan, UT
Abstract
Following the success of the first Phi-Sat mission, in 2020, the European Space Agency (ESA) announced the opportunity to present CubeSat-based ideas for the Phi-Sat-2 mission as part of its initiative to promote the development of radically innovative technologies such as Artificial Intelligence (AI) capabilities onboard Earth Observation (EO) missions. Open Cosmos and CGI submitted a joint proposal for the Phi-Sat-2 mission idea, which takes advantage of the latest research and developments in the European ecosystem. The proposed mission idea is a game-changing EO CubeSat capable of running AI Apps that can be developed, easily deployed on the spacecraft and updated during flight operations. The AI Apps can be operated from ground using a simple user interface. This approach allows continuous improvement of the AI model parameters using the very same images acquired by the satellite. The mission takes advantage of the latest research for mission operations of CubeSats and use the NanoSat MO Framework, a framework for small satellites that allows software to be deployed in space as simple Apps, in a similar fashion to Android apps. This framework was previously demonstrated in ESA’s OPS-SAT mission, and supports the orchestration of on-board Apps. It fully decouples the App features from the underlying on-board hardware via an abstraction layer API in the form of services. Additionally, it includes a Software Development Kit with demo Apps, development tools, and tutorials to facilitate the development of Apps. By decoupling the data platform from the Apps, it is possible to distribute the development of specialized AI Apps to different partners within the Phi-Sat-2 mission consortium. The mission will include a set of default AI Apps that will be able to do vessel detection, forest monitoring, and roadmap transformation from satellite imagery. The framework allows more than just the set of default Apps and so, third-party Apps can be included on later stages of the mission lifecycle. This paper will present the NanoSat MO Framework, introduce the AI Apps that are part of the Phi-Sat-2 mission, and how the free and open-source framework enables the creation of software-defined satellite missions via on-board Apps.
NanoSat MO Framework: Enabling AI Apps for Earth Observation
Utah State University, Logan, UT
Following the success of the first Phi-Sat mission, in 2020, the European Space Agency (ESA) announced the opportunity to present CubeSat-based ideas for the Phi-Sat-2 mission as part of its initiative to promote the development of radically innovative technologies such as Artificial Intelligence (AI) capabilities onboard Earth Observation (EO) missions. Open Cosmos and CGI submitted a joint proposal for the Phi-Sat-2 mission idea, which takes advantage of the latest research and developments in the European ecosystem. The proposed mission idea is a game-changing EO CubeSat capable of running AI Apps that can be developed, easily deployed on the spacecraft and updated during flight operations. The AI Apps can be operated from ground using a simple user interface. This approach allows continuous improvement of the AI model parameters using the very same images acquired by the satellite. The mission takes advantage of the latest research for mission operations of CubeSats and use the NanoSat MO Framework, a framework for small satellites that allows software to be deployed in space as simple Apps, in a similar fashion to Android apps. This framework was previously demonstrated in ESA’s OPS-SAT mission, and supports the orchestration of on-board Apps. It fully decouples the App features from the underlying on-board hardware via an abstraction layer API in the form of services. Additionally, it includes a Software Development Kit with demo Apps, development tools, and tutorials to facilitate the development of Apps. By decoupling the data platform from the Apps, it is possible to distribute the development of specialized AI Apps to different partners within the Phi-Sat-2 mission consortium. The mission will include a set of default AI Apps that will be able to do vessel detection, forest monitoring, and roadmap transformation from satellite imagery. The framework allows more than just the set of default Apps and so, third-party Apps can be included on later stages of the mission lifecycle. This paper will present the NanoSat MO Framework, introduce the AI Apps that are part of the Phi-Sat-2 mission, and how the free and open-source framework enables the creation of software-defined satellite missions via on-board Apps.
