Parameterized fuzzy-logic controllers for the attitude control of nanosatellites in low earth orbits. A comparative studio with PID controllers

Abstract

The pointing requirements for a satellite (e.g. for its payload or for its subsystems) determine the different attitude control system operational modes that are necessary along its mission. Two examples are: a low cost mode for nominal attitude stabilization; and a faster mode for antenna pointing during communication phases. In this work, an attitude control algorithm based on fuzzy logic is presented to achieve an optimal performance of the controller for different operational requirements of the satellite mission. After analysing the design process and the response of previous fuzzy controllers, several improvements have been implemented concerning the number of membership functions, their shape, and the contents of the fuzzy rules, obtaining a parameterized final fuzzy inference system whose performance depends on eighteen parameters, six per each of the three control axes. The design proposed in this work will be tested soon on board the European Space Agency OPS-SAT mission and it is easily extensible to most nanosatellite missions in Low Earth Orbits.Different sets of the eighteen tuning parameters have been used to analyse the design and the performance of this new fuzzy controller. The values have been selected through multi-objective genetic optimizations to deal with two conflicting objectives: low cost and accuracy. A numerical performance comparison shows that the new design provides more efficient controllers than those used in previous works. In addition, the performance of the new fuzzy controller has been evaluated against traditional Proportional Integrative Derivative (PID) controllers for different satellite operational requirements. The values of the PID gains have also been selected through the same multi-objective genetic optimization. The comparisons show that the new fuzzy design provides great advantages over the PID strategy in most of the operational modes considered, and how the PID strategy has difficulties to produce a low cost control law profile. Furthermore, the new fuzzy design is able to improve the Pareto front to optimal points that are unreachable with the PID schema. These results open new possibilities for the application of intelligent controllers in the space industry.