Modeling and control of color tunable lighting systems

[Display omitted] •Experimental validation of an optimization-based controller for smart lighting.•Trade off light quality and energy usage by suitable choice of cost functions.•Human comfort included through empirical factors such as the Kruithof curve.•Online model adaptation from input–output data to improve control performance. Electric lighting has not substantially changed in over 100 years. From incandescent bulbs to fluorescent tubes, the efficiency remains low and control mostly involves on/off or dimming. The new wave of solid state lighting offers the possibility of sensor-based intensity regulation, color control, and energy efficiency, under varying needs and environmental conditions. This paper formulates the lighting control problem as an optimization problem balancing color fidelity, human perception and comfort, light field uniformity, and energy efficiency. The optimization problem is solved based on the light propagation model, which is adaptively updated with color sensor feedback to account for changing ambient lighting conditions, such as daylighting. We demonstrate the proposed approach in a smart space testbed under a variety of use conditions. The testbed is instrumented with 12 color tunable lights and 12 light sensors, as well as simulated daylight. The results show substantial improvement in terms of energy usage and delivering good light field quality in the presence of varying lighting conditions. Experimental results corroborate the efficacy of the proposed algorithms.