Attitude and orbital determination of a picosatellite : a testing platform design

Abstract

The determination of a spacecraft’s orientation (attitude) and orbital position is a crucial task that needs to be constantly performed during the spacecraft’s lifetime. This is done to ensure that communication from earth to the spacecraft or satellite is maintained and therefore orbital commands can be sent or experimental data can be downloaded. The UoMBSat-1 is a picosatellite being designed by a team of researchers ASTREA, at the University of Malta. The picosatellite is restricted by the PocketQube standard; ie. having dimensions 5 × 5 × 5 cm and weighing 250 grams at most. One of its main systems that are being designed is the Attitude Determination and Control System (ADCS). This system requires to be tested using a conical pendulum that simulates the satille in orbit; which is a method that has never been proposed before in picosatellite applications. The project designed in this thesis is a hardware prototype that serves as a platform where an attitude and orbital determination algorithm can easily be implemented on. The prototype’s function in this thesis is to acquire and process the data from three on-board sensors and then wirelessly send the data to a computer terminal to be evaluated. One of the main aspects of this thesis is the design of a 27-state Finite State Machine (FSM) to be used with the sun sensors. The FSM is used to obtain a logical sequence of motion of the satellite based on the direction of the sun (sun vector). All sensor outputs are processed such that they are ready to be fed through an attitude estimation algorithm when the AODS is ready to tested on the conical pendulum. In this project, a successful operation of the testing platform hardware was confirmed with all sensors generating data in the format required by the estimation algorithm (quaternions). The FSM transition were confirmed and then implemented to be used with the sun sensors. Testing of the implemented FSM was performed and a sun vector angle error of 6.4% was obtained. A dead-reckoning orbital estimation algorithm was also implemented to be used with the AODS in the future.