A Machine Learning Approach for Global Steering Control Moment Gyroscope Clusters

Abstract

This paper addresses the problem of singularity avoidance for a 4-Control Moment Gyroscope (CMG) pyramid cluster, as used for the attitude control of a satellite using machine learning (ML) techniques. A data-set, generated using a heuristic algorithm, relates the initial gimbal configuration and the desired maneuver—inputs—to a number of null space motions the gimbals have to execute—output. Two ML techniques—Deep Neural Network (DNN) and Random Forest Classifier (RFC)—are utilized to predict the required null motion for trajectories that are not included in the training set. The principal advantage of this approach is the exploitation of global information gathered from the whole maneuver compared to conventional steering laws that consider only some local information, near the current gimbal configuration for optimization and are prone to local extrema. The data-set generation and the predictions of the ML systems can be made offline, so no further calculations are needed on board, providing the possibility to inspect the way the system responds to any commanded maneuver before its execution. The RFC technique demonstrates enhanced accuracy for the test data compared to the DNN, validating that it is possible to correctly predict the null motion even for maneuvers that are not included in the training data.

Ecosystem of Aviation Maintenance: Transition from Aircraft Health Monitoring to Health Management Based on IoT and AI Synergy

“…The study [65] brings forward machine learning techniques, particularly deep neural networks, and random forest classifiers, for predicting null motions in control moment gyroscopes, which are pivotal for satellite attitude management. This approach exhibits remarkable accuracy, making it a reliable tool for anticipating null motions beyond the initial training scenarios.…”

Dynamic analysis of a rigid-flexible inflatable space structure coupled with control moment gyroscopes

“…FF Φ λ F F 0 Kq (28) By solving Eq. ( 28) with the Newton-Raphson iteration method, the generalized coordinate 0 q of the equivalent static equilibrium position can be obtained.…”

“…Frederick [27] derived the equations of motion from first principles for a very general case of a spacecraft with n-CMG, and described each contribution of the dynamics equation and its effect on CMG performance. Papakonstantinou et al [28] dealt with the problem of singularity avoidance for a 4-CMG pyramid cluster by using the attitude control of a satellite using machine learning techniques. Indeitsev et al [29] presented the results of modeling the nonlinear dynamics of vibrating wheel gyroscope.…”

Artificial Intelligence in Aviation: New Professionals for New Technologies

“…The study [27] introduces machine learning techniques, specifically Deep Neural Network (DNN) and Random Forest Classifier (RFC), to predict null motions in a 4-Control Moment Gyroscope used for satellite attitude control. The RFC method shows superior accuracy over the DNN, allowing for reliable predictions of null motions even for maneuvers not in the training data.…”

“…The frequency and links diagram in Figure 5 gives a link analysis of the first 12 words in the corpus. This diagram, for example, shows links for the words engineering (the most frequent word in the corpus with a frequency of 387) with links to aircraft (190 links), students (152), knowledge (39), mechanics (29), and aerospace (27).…”