QuickSAT/Autonomy - an open Autonomy Framework for Cubesats and Small Satellites

Author:

Abstract
The QuickSAT/Autonomy is a framework of autonomy APIs and system with supporting flight computer designed for cubesats to support Tactical Intelligence, Surveillance, and Reconnaissance (ISR) functions plus generic autonomy related functions. The system allows the cubesat through a range of sensors such as a hyperspectral camera, which was used in our test to collect data, to “perceive” key select information, and alert to the ground what is “sensed”, relaying only the required critical data. Operators can later download larger data sets as needed. This will allow the streamlining what is communicated to the ground in a timely manner. Communications is a critical bottleneck in small satellites such as cubesats – on board autonomy can rapidly assist in ISR functions. Distributed systems using formation flying satellites can use synergy of autonomous payloads on­board of different satellites instead of a multiplying effect of constellations in order to enhance coverage. The open framework allows the QuickSAT/Autonomy System with the LinkStar avionics system is to provide a flexible vehicle management system, a range of autonomy tools, global communications, and accurate and rapid tracking of 1U and larger cubesats and small satellites, providing exceptional situational awareness and ISR support. Messages are received by the ground within 200ms of transmission from over 95% of Low Earth Orbit. The heart of the LinkStar avionics system is the Xilinx Zynq UltraScale+ MPSoC chip, a system that combines a quad core ARM-53 processor and Zynq-7000 FPGA. The architecture however is flexible to allow it to be scaled to smaller, less power consuming processors. LinkStar supports I2C, I2C multiplexer, Space Wire, serial, USB and CAN allowing for a range of sensors to be connected to the QuickSAT/Autonomy architecture; other data buses can be easily infused into the design. LinkStar also includes a power management software, software rad-hardening tools, and QuickSAT/Xen hypervisor to enable multiple, secure operating systems on the processor. The board design is a PC104 format with full support for the Cubesat bus. The integrated flight computer hosts the QuickSAT/VMS system providing vehicle control, communications, and instrument management functions. QuickSAT/VMS provides a web based interface for easy vehicle configuration, system testing, and management. For the this presentation we will present the AI framework and the overall architecture, deep learning mechanisms implemented in the framework, and share test results. In this presentation we will also share how to download and implement the software, and a example of identification of a commercial aircraft from LEO vie multispectral imaging sessions. We will also show the architecture is being used in the CaNOP/EarthScan Mission in November 2020 and by our partner organizations.