Adaptive Sliding Mode Control for Fast Steering Mirror Based on RBF Neural Network Self-Learning

The fast steering mirror (FSM) is a key element in electro-optical systems. For designed fast steering mirror driven by voice coil motors, considering the plant parameter uncertainty due to the unmodeled dynamics, an adaptive sliding mode controller with a robust term constructed based on the RBF neural network self-learning is proposed. The system identification indicates that RBF neural network (RBFNN) can achieve good approximation to the system parameters. Considering the output limitation, the input signal range is determined through the joint simulation of Adams and Matlab. The numerical simulation verifies that the proposed control scheme can achieve high-precision tracking performance. At the same time, the tests in two cases without input constraint and with input constraint are conducted. The test data illustrates that the developed control scheme with input constraints can better protect the FSM system.