Resource-constrained avionics design for CubeSats

Abstract

We present an optimization approach to CubeSat avionics design which considers the consumption of some resources (electrical power, volume) and production of others (processing power, volatile memory, non-volatile memory, and radiation tolerance) in a quantitative optimization analysis. We present the avionics hardware design for the Microwave Radiometer Technology Acceleration (MiRaTA) 3U CubeSat, funded by the NASA Earth Science Technology Office (ESTO), as a case study for our optimization analysis. MiRaTA will demonstrate a three-band microwave radiometer and GPS radio occultation (GPSRO) sensor suite for profiling atmospheric temperature, humidity, and cloud ice. The goal is to increase the Technology Readiness Level (TRL) of the weather-sensing technology from TRL 5 to TRL 71. The avionics system is the "central nervous system" of the spacecraft, managing interfaces with every subsystem and between the Bus and Payload. MiRaTA's avionics design supports the Payload, which is tasked with the science mission to gather and process appropriate radiometer and GPSRO data, and the Bus, which comprises subsystems to handle attitude determination and control (ADC), power regulation and distribution, communications with the ground station, thermal management, and a suite of sensors and telemetry components. MiRaTA's avionics system uses a custom designed motherboard with a PIC24FJ256GB210 microcontroller to command activity in the Bus and manage data and power for the Payload. This custom Motherboard - dubbed the "Micron Motherboard" - leverages many of the advantages of the popular Pumpkin Motherboard but with reduced complexity and improved performance. The MiRaTA avionics system is also designed to minimize the number and length of cables, simplify connector uniformity, and improve accessibility. The design improvement in avionics hardware from MicroMAS to MiRaTA is quantified using an optimization coefficient: 1.522. We expect optimization coefficients to range typically from -4 to +4, so this design indicates a modest improvement.