Time difference simultaneous perturbation for neurocontrol

This paper proposes a neurocontroller via the time difference simultaneous perturbation learning rule to control an unknown plant. When we apply a direct inverse control scheme by a neural network, the neural network must learn an inverse system of the unknown plant. Therefore, we must know the Jacobian of the plant, when we use a kind of gradient method as a learning rule of the neural network. On the other hand, our control scheme described here does not require information about the plant Jacobian because the time difference simultaneous perturbation method estimates the gradient by using a kind of the finite difference. A tracking problem for a dynamic plant is shown to confirm the feasibility of the method