On the momentum term in gradient descent learning algorithms
Neural Networks
Convergence of exponentiated gradient algorithms
IEEE Transactions on Signal Processing
| Hi-index | 0.00 |
A new learning algorithm for fuzzy system to approximate unknown nonlinear continuous functions is presented. Fast terminal sliding mode combining the finite time convergent property of terminal attractor and exponential convergent property of linear system is introduced into the conventional back-propagation learning algorithm to improve approximation ability. The Lyapunov stability analysis guarantees that the approximation is stable and converges to the unknown function with improved speed. The proposed fuzzy approximator is then applied in the control of an unstable nonlinear system. Simulation results demonstrate that the proposed method is better than conventional method in approximation and tracing control of nonlinear dynamic system.