Optimizing LED Driver Transients Using ANN and ANFIS Control in SEPIC Converters

In modern lighting systems, where reliable electrical regulation directly affects performance and longevity, the stability and efficiency of LED drivers are essential. Most LED drivers prefer Single-Ended Primary Inductor Converters (SEPIC) because they can handle both step-up and step-down voltage requirements. Besides, PID and traditional control techniques can not properly and adequately handle the dynamic and nonlinear behavior of these converters; which results in inefficiencies and suboptimal control under a variety of conditions. This paper introduces a novel approach utilizing Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS) to enhance the control of SEPIC converters in LED drivers. Accordingly, these smart control methods allow dynamic adaptation to system changes with enhanced voltage stability, transient response, and energy efficiency. This work, based on extensive simulations and comparisons of performance, demonstrates that the controllers based on ANN and ANFIS convincingly outperform the traditional ones while opening new perspectives toward more resilient and energy-efficient LED driver systems.