Turner and Hall Adaptive Spacecraft Attitude Control ADAPTIVE SPACECRAFT ATTITUDE CONTROL USING NEURAL NETWORKS

Neural networks offer a unique approach to controlling a dynamically changing system. By updating synaptic weights of an interconnected, algebraic framework, desired system objectives can be reached despite an unknown operating environment. This research implements a neural network for controlling the attitude of a satellite with unknown system disturbances. The training and updating of the controller occurs on-line in real-time, increasingly minimizing pointing error during mission operations. The design of the neural network controller is discussed, as well as preliminary results from computer simulations. Introduction The majority of current satellite control technologies employ classical methods that are determined a priori. However, due to changes in the satellite system or unknown flight characteristics, these systems may prove to be unreliable. Adaptive control systems provide a robust means of controlling a satellite in the face of these uncertainties and failures. In this research project we develop a control system that adapts to the satellite’s flight characteristics. By using ‘soft-computing’ techniques such as a neural network, the control system learns and adapts to changes in the controllability of the satellite. Using existing methods of neural network development [1], an adequate controller can be designed that will control the three axis angles, and three rotation rates of a satellite’s attitude [2][3][4][5]. The use of an adaptive attitude control system can greatly reduce the costs of satellite algorithm and software development. An adaptive control system can learn from a model of the satellite and from the actual flight mission instead of having to specify the classical control law. Furthermore, a well-designed neural network can be reused for many missions, without having to change the base structure of the system. This reuse can greatly reduce the cost of having to develop a control system from the ground up. Also, the adaptive control system can recover from system failures that may otherwise cripple a spacecraft such as those that occurred on Mariner 2 and PanAmSat Corporation's Galaxy 4 [6]. The robust satellite can recover from these failures without having to wait for human intervention and updating of its software. Spacecraft Model In order to design a controller for spacecraft attitude, a model of the system must be determined. The nonlinear spacecraft model includes the spacecraft attitude kinematics, rotational dynamics and control torques. It is assumed that the spacecraft is measured in bodyorbital coordinates, with a nominal nadir pointing attitude. That is, the body axes are initially aligned with the orbital axes, and the body z-axis points towards the earth. The 4-element quaternion set, q , is used to specify the attitude, as there are no singularities similar to those found in an Euler angle set. The quaternion can be determined from the Euleraxis parameters set (a, φ) as follows: