Universal approximation using radial-basis-function networks
Neural Computation
Model-based iterative learning control with a quadratic criterion for time-varying linear systems
Automatica (Journal of IFAC)
Hi-index | 0.00 |
In this paper, a run-to-run control with neuro-fuzzy model updating mechanism is developed. This strategy features the ability to learn from previous batches to obtain iteratively the optimal control profile and adjust the neurofuzzy model parameters. In addition, an updating algorithm guaranteeing the global convergence of the weights of the model is developed based on the Lyapunov approach. As a result, model uncertainties can be handled. Simulation results show that by updating the model from batch to batch, the control profile converges to the corresponding suboptimal one in the subsequent batches.