Attitude Control and VLS-1 Rocket Public Benchmark Model

Codes and guidance data related to the work "Case Study on PID-based Attitude Control of the VLS-1 Rocket Using a Publicly Available Model".

MATLAB codes related to the following work:

Title: Case Study on PID-based Attitude Control of the VLS-1 Rocket Using a Publicly Available Model.

Abstract: In this work, three Proportional-Integral-Derivative (PID) digital control syntheses were investigated in terms of robustness and performance assessment in the attitude control of a satellite launch vehicle based on the Brazilian VLS-1 rocket during the transonic and maximum dynamic pressure phases. The three digital control syntheses were obtained based on Euler's backward difference approximation, rearranging the PID terms to assess the parallel PID, the I-PD, and the PI-D control-loop topologies. These controllers were tuned by a linear quadratic method, exploiting its optimality. The stability analysis was investigated using classical frequency response techniques, highlighting important remarks for their use with open-loop unstable systems. The robustness of the designed controllers to modeling uncertainties and output disturbances was quantified based on gain and phase margins obtained from the amplitude spectrum of the stabilized closed-loop sensitivity and complementary sensitivity functions. The designed attitude control systems were also studied in the time domain, considering the simulation with a 16th-order model, asymmetrical and coupled, with the presence of bending modes, non-collocated actuator-sensor dynamics, sensor noise, wind as low-frequency disturbance, and non-linear input saturation. The studied controllers were compared to a baseline PI-D designed for the unfinished fourth version of the VLS-1 rocket. The main results were obtained from four sets of experiments, dedicated to select the best PID structure after successive performance and robustness analyses, energy and power efficiency evaluation, and adherence to design specification requirements. Beyond these large amount of experimental results in different and defying control scenarios, this work also contributes by sharing the used Matlab-based VLS-1 model, making it available for download and use by the general public for future research and enhancements.