ESA GNC Conference Papers Repository

Adaptive Control for Vibration Attenuation of a Laser Communication Terminal
Frederik Thiele, Iñigo Fernandez Imaña, Xavier Manuel Juanpere, Harald Pfifer
Presented at:
Sopot 2023
Full paper:

Laser Communication is the next generation of communication technology in the space industry, finding use in near earth as well as deep space application. They provide secure connections with significant higher data rates and less power consumption compared to conventional methods. Moreover, they address the ongoing problem of frequency saturation in radio communication technology. A challenge for optical systems is the high pointing requirement necessary to establish reliable communication links, which increase with higher data rates or distance. In this work a control method using estimation and adaptive algorithms is presented to attenuate vibrations as the main contributors for depointing in a standard Laser Communication Terminal (LCT) setup. Higher tracking performance compared to a baseline controller is achieved and the potential for decreasing hardware requirements, and thus enabling cost reduction, is analysed. The investigated standard LCT has a rigid telescope and focal plane assembly. A Fine Pointing Mechanism (FPM), which shows flexible behaviour, provides pointing angle adjustments, while an Acquisition and Tracking Sensor (ATS) determines the depointing measurement based on the transmitted and received signal. The entire terminal is connected to the satellite with a flexible mounting structure, which includes a Coarse Pointing Mechanism (CPM) to change the telescope's attitude. Establishing and maintaining an optical communication link requires a complex sequence and multiple LCT operational modes. After an initial acquisition and alignment phase the data transmission is conducted, which imposes the highest constraints for performance and robustness on the system. Those guarantee a continuous communication link and ensure specified data rates. Therefore, only the transmission mode is investigated within this paper. During this operational mode the CPM holds its position and can be considered quasi-static. The pointing adjustment is provided by the FPM in closed-loop operation. The system experiences several external disturbances, which can deteriorate pointing performance. They are transferred from the satellite to the terminal through the CPM. Typical sources are thermoelastic behaviour, solar array rotation and various micro-vibrations (MVs). Thereof MVs created by the reaction wheels for the satellite’s attitude control are identified as the main contributor, as their frequencies can coincide with the CPM-Modes. Hence only the impact of the reaction wheels is considered within the presented study. At any given time, multiple superimposed MVs from the reaction wheels are present within the system. They can each be modelled as a sinusoid with variable frequency and amplitude. After CPM transfer, they act on the closed-loop as an output disturbance during transmission mode. The general envelope for frequencies and amplitudes is known due to the reaction wheels specifications. However, the exact values at a specific time are considered unknown as no communication between the satellite and the LCT is assumed. The CPM orientation and thus its mode is indefinite as well. Only the FPM and ATS transfers are considered known. The required pointing performance can be provided using high bandwidth controllers and high sample rates, needing expensive hardware for their implementation. Within this work an alternative control method using estimation and adaptive algorithms is presented. It is based on the principle of destructive interference and particularly suited to attenuate sinusoidal disturbances with the potential to decrease hardware requirements compared to high bandwidth controllers. An observer is defined to provide an estimation of the disturbances imposed on the system after CPM transfer. It uses the residual disturbance measurement by the ATS as well as the ATS and FPM transfer. The synthesized control signal for the FPM is also fed back into the observer, allowing the estimation of the original disturbance despite attenuation. The observer provides an estimation of the disturbance in its entirety, however synthesizing the appropriate control input requires the specific properties, such as frequency, amplitude and phase-offset of each present sinusoid separately. A Fast Fourier Transformation (FFT) of a discrete signal section and analysis of the obtained frequency spectrum provides the current frequency of the most dominant sinusoid in the disturbance. Then a parameterized signal is created adaptively, which matches the amplitude and phase-offset of the disturbance part corresponding to the frequency provided by the FFT. The implemented gradient algorithm uses the solution of two differential equations minimizing a cost function, which represents the residual disturbance. A cascade setup of the described method is used to handle all parts of the disturbance. An intermediate signal processing step is added to prevent duplication in the frequency estimation by removing already identified signals from the further investigated estimated disturbance. In a concluding step the parameters are adjusted according to the FPM transfer and an additional phase-offset of 180° is introduced to synthesize the necessary control signal to attenuates the disturbance at FPM output. A nonlinear simulation is used to evaluate the pointing performance for a realistic scenario with multiple MVs. A Monte-Carlo study for several CPM-Modes, which also includes sensor noise and FPM transfer uncertainty, showed a significant improvement in pointing performance, especially during CPM resonance, compared to a simple high bandwidth controller provided as a baseline reference by Airbus. Moreover, preliminary investigations about the impact of different sample rates are conducted. The presented work was developed in cooperation between the Technische Universität Dresden and Airbus Defence and Space as part of the ESA-funded project No. AO/111017/21/NL/MGu entitled Adaptable Control and Estimation with Guaranteed Robust Performance (ACE).