Learning control for time-delay systems with iteration-varying uncertainty: a Smith predictor-based approach | Synapse