2019.A.2.4. System design concept of Mars Intelligent Imaging & Atmospheric Research CubeSat Constellation using Distributed Deep Learning (MIIAR)

Author(s)

Rashedul Huq (1)
Monirul Islam (1)
Partha Pratim Das (1)
Shahnewaz Siddique (1)

North South University, Bangladesh

Session

A.2

Keywords

Hyperspectral Imaging, IR emission spectroscope, Distributed Deep Learning, Inter-planetary mission, Software Defined Radio

Abstract

Planetary observation nanosatellites are increasingly popular since they can be deployed quickly and inexpensively. In this paper, we have designed an innovative and intelligent planetary mission for imaging targeted locations and atmospheric study on Mars from a constellation of nanosatellites. The MIIAR constellation consists of two sets of total seven CubeSats, each for a specific task and function and each set deployed in a specific orbit. The first set is a data relay CubeSat, named as Mothership Data Relay Satellite (MS-DRS), containing three 12U units deployed in the aero-stationary orbit of Mars, whose principal mission is to relay data to Earth from the observational CubeSat deployed on Mars. In addition, the Daughter Ship Type-I CubeSat (DS-I) and Daughter Ship Type-II (DS-II) is a cluster of four 6U observational satellites, deployed in the high inclined orbit of Mars, containing dedicated instruments, for research and study of Mars’ atmosphere and hyperspectral imaging of its surfaces. Another important module, a low powered Deep Learning accelerator is included for the first time in a real-time system (to our best knowledge) to perform various machine learning and deep learning operations on board when required and shared among the constellation, creating a Distributed Deep Learning platform among the constellation. For advanced communication, we have included a Software Defined Radio (SDR) and an optical wireless communication module that will provide an advanced inter-satellite link and a deep space communication system for the constellation, thus, reducing the sat-failure probability. Through MIIAR, the scientific community can gather detailed information for future human colonization missions on Mars.