An Accurate, Inexpensive Attitude Determination and Control System for CubeSats by Duarte Rondão, Cass Hussmann and Afzal Suleman

From The Space Library

Jump to: navigation, search

Duarte Rondão*, Instituto Superior Técnico, University of Lisbon, Portugal (duarte.rondao@tecnico.ulisboa.pt)

Cass Hussmann Centre for Aerospace Research, University of Victoria, Canada (hussmann@uvic.ca)

Afzal Suleman,Professor and Canada Research Chair, Department of Mechanical Engineering, Centre for Aerospace Research, University of Victoria, Canada (suleman@uvic.ca)

Contents

Abstract

Attitude determination and pointing control in cube satellites has been a continuous challenge mainly due to volumetric constraints and lack of small attitude sensors. In addition, the attitude control systems of CubeSats typically employ magnetorquers as the main actuators, which have a full range of control at polar orbits, but become less effective at lower inclinations. They also suffer from reduced pointing accuracy when compared to reaction wheels. As a result, most CubeSat control designs are characterized by considerable attitude errors. This paper proposes an attitude determination and control system (ADCS) using low-cost sensors and actuators that allows for an ADCS with an accuracy better than 4.78 deg. For attitude determination (AD), the development and comparison of four quaternion-based algorithms from different families of estimators is performed using vector measurements from a sun sensor and magnetometer, and angular velocity measurements from MEMS gyroscopes. The notable Multiplicative Extended Kalman Filter (MEKF) is implemented as a benchmark for the other three algorithms: the Unscented Quaternion Kalman Filter (UQKF), the Constant Gain GEometric Attitude OBserver (GEOB) and the Two-Step Optimal Estimator (2STEP). All algorithms are designed for the estimation of the quaternion of rotation and the gyro bias. For attitude control (AC), an enhanced version of the B-dot control law using magnetic torquers is implemented for the detumbling stage. A 3-axis sliding mode controller is employed for the nominal, Earth-tracking mission phase, using a specially designed reaction wheel system (RWS). A momentum dumping system using the magnetic torquers is also devised to ensure that the RWS does not reach saturation. Simulations are performed in an environment for the ECOSat-III cube satellite: a nanosatellite designed by the University of Victoria student team who achieved first place in the 2014 Canadian Satellite Design Challenge (CSDC). These tests comparatively demonstrate the attainable performances of the four AD methods, allowing for the selection of the most adequate one for the ECOSat mission scenario. The robustness of the AD algorithms is evaluated. The results demonstrate that an accurate ADCS for CubeSats is obtainable on a restricted budget, without resorting to high precision optical sensor suites.


I. Introduction

A CubeSat (cube satellite) is a type of miniaturized satellite formed from the combination of 10 cm cubes, each with a mass of up to 1.33 kg. The CubeSat project was born in 1999 from a joint effort between Cal Poly and Stanford University to promote an inexpensive way for universities and students to develop skills and experience on the design, manufacturing and testing of small satellites. From this outline, it becomes evident that CubeSats are subjected to strict cost, power, mass and volume constraints. A typical ADCS, in particular, constitutes a very large percentage in terms of these four budgets for a regular satellite. Therefore, classical attitude determination and control strategies must not be expected to work on a CubeSat without some adaptation, which by itself is not an easy task. This leads to cube satellite designs characterized by poor pointing accuracies. This market gap has been seen as an opportunity for some, and everyday more commercial-off-the-shelf (COTS) attitude hardware is becoming available for the regular consumer. In addition, it is also possible to do an in-house customized manufacture of these hardware components. This paper proposes an ADCS capable of delivering a pointing performance better than 4.78 deg. The system is designed in the framework of the ECOSat-III satellite, the second satellite project of the University of Victoria within the Canadian Satellite Design Challenge (CSDC). During the course of its mission, ECOSat-III will need to fulfill certain pointing requirements. The initial task of the ADCS is to perform the detumbling of the satellite after separation. Then, it must be able to adjust its attitude from an arbitrary orientation in order to align its hyperspectral camera with nadir. This alignment is smoothly maintained throughout the orbit in what is called Earth-tracking (or nadir-tracking) mode. During passes over Victoria, BC, the satellite may be required to enter targeting mode to point its antenna directly to the ground station. It may also be required to enter safe mode, where the solar panels are aligned with the Sun’s direction to maximize power generation. The ADCS of ECOSat-III features a rate gyro, a magnetometer, a sun sensor, six magnetorquer rods and four reaction wheels in order to achieve the pointing requirements. The structure of this paper is as follows. For the remainder of Section I, additional insight is given on the CSDC and the ECOSat team. Section II comprises the mathematical bases required to develop the attitude estimation algorithms in Section III and the attitude control algorithms in Section IV. In Section V the choice of sensors and actuators is discussed. In Section VI, numerical simulations for a realistic ECOSat-III orbit scenario are performed. Section VII draws conclusions on the analysis performed. Section VIII presents future work recommendations for the project.

A. The Canadian Satellite Design Challenge

The Canadian Satellite Design Challenge is a competition spanning 12 Canadian universities consisting in the development and design of a 3U (10 × 10 × 30 cm) cube satellite and the scientific missions it will conduct. The challenge promotes the creation of small satellite infrastructures, as well as knowledge and research into the practical and commercial applications of nanosatellite technologies in Canada. The teams must go through preliminary and critical design reviews, as well as vibration testing of their prototype, which are then judged with the first place winning assistance in launching their satellite. A large focus of the CSDC is to reach out to the community and students to promote science, engineering, and space as exciting and viable career paths as well as to promote Canada’s place in space technology. The created nanosatellite prototype for the competition has to have all the critical functionality of a larger satellite such as the power, payload, attitude determination and control, communication, and operations.

B. The University of Victoria ECOSat Team

The University of Victoria (UVIC) ECOSat Team is a student group who has competed in the CSDC since 2011. It is comprised of a diverse group of graduate and undergraduate students, passionate about space technology and technological innovation in general. The initial satellite developed by UVIC placed third in the national competition in 2012. Furthermore, ECOSat proudly achieved first place in 2014 with the ECOSat-II CubeSat, securing a launch into low Earth orbit. Notably, rather than purchasing existing hardware and adopting third party software, the command and data handling systems, mechanical construction, payload development, and power system have been designed and developed primarily in-house at UVIC.

The current team project is the ECOSat-III CubeSat (Figure 1). ECOSat-III will further UVIC’s contribution to the geophysical service and to research and development of communication systems on nanosatellite systems. The satellite will be flying a primary hyperspectral imaging payload, supported by an experimental communications system and attitude control system. The mission objectives are: A) to provide hyperspectral imagery of Canada at 150 metre resolution; B) to downlink the hyperspectral imagery over a custom-developed 40 MBit communications system; C) to extend the UVIC’s experience in attitude determination and control systems with the addition of momentum wheels and more complex attitude determination algorithms; D) to provide accurate initial orbit determination and low rate telemetry through the use of an experimental below-the-noise-floor communications system.

Please continue to Part 2 here