Model Predictive Control for Luminous Flux Tracking in Light-Emitting Diodes

Luminous flux tracking and junction temperature stress minimization are typical objectives in the regulation of high-brightness light-emitting diodes (LEDs). In this brief, a solution based on a model predictive controller approach is proposed. The state estimation is obtained using an LED photoelectrothermal dynamic model whose parameters are tuned through datasheets and dedicated experiments. By exploiting the dynamic separation within the thermal state variables of the model, the junction temperature is predicted using LED current and heat sink temperature measurements. The effectiveness of the proposed approach is verified through simulations and experiments on different heat sinks and samples of the same LED family, thus confirming the achievement of the luminous flux tracking under current and temperature constraints typically featuring LED applications.