Phi-Sat-2 (also known as ɸ-Sat-2) is an Earth observation CubeSat mission from the European Space Agency (ESA) capable of running AI apps directly on board. What makes Phi-Sat-2 particularly noteworthy is its utilization of the NanoSat MO Framework, a modular and open-source platform designed for small satellite missions.[1]
The NanoSat MO Framework enhances the satellite's flexibility and adaptability, allowing for efficient integration of AI technologies into its operational workflow through Apps that can be installed on board. The AI Apps will be able to do different activities such as transforming a satellite image to a street map, detecting clouds, detect and classify maritime vessels, and to perform image compression using AI.[1][2]
Mission Consortium
The ɸ-Sat-2 mission consortium is composed of the following companies:
The mission includes the following payload devices:
On-board Computer from Open Cosmos
Multi-spectral Optical Camera from Simera CH Innovative (expected 4.75 m ground resolution)
AI processor: Intel Movidius Myriad 2 from Ubotica[5][1]
The AI technology used in the mission is based on the Intel Movidius Myriad 2 vision processing unit, which is designed to provide high-performance, low-power processing for computer vision applications. The Myriad 2 is integrated into the spacecraft and is used to process and analyze the images captured by the hyperspectral camera in near real-time.
The AI processor was already adopted on the previous Phi-Sat-1 mission.[5]
Launches are separated by dots ( • ), payloads by commas ( , ), multiple names for the same satellite by slashes ( / ). Crewed flights are underlined. Launch failures are marked with the † sign. Payloads deployed from other spacecraft are (enclosed in parentheses).