Neural Network Augmented Adaptive Backstepping Control of Spacecraft Proximity Operations

Liang Sun,Bing Zhu
DOI: https://doi.org/10.1007/978-981-32-9682-4_22
2019-01-01
Abstract:This paper investigates the six degrees-of-freedom relative motion control of spacecraft proximity operations under multiple uncertainties. A radial-basis-function neural network (RBFNN) adaptive backstepping controller is developed. The relative dynamics controller is developed by backstepping technique, and model uncertainties are compensated by RBFNNs. Each RBFNN utilizes a smoothly second-order piecewise function to combine adaptive control and RBFNN. The RBFNN holds the response performance in neural active region, and adaptive control dominates response performance outside the active region. It is proved that the proposed control method can guarantee the relative pose is ultimately uniformly bounded. Simulation demonstrates the performance of the control method.
What problem does this paper attempt to address?