Optimal LED-based illumination control via distributed convex optimization

Achieving illumination and energy consumption targets is essential in indoor lighting design. The provision of localized illumination to occupants, and the utilization of natural light and energy-efficient light-emitting diode (LED) luminaires can help meet both objectives. Localized illumination control schemes require suitable coordination mechanisms to obtain luminaire dimming levels that achieve the desired illuminance levels and reduce energy costs. This paper presents several distributed optimal LED-based illumination control schemes that provide localized illuminance to occupants. The lighting system consists of multiple LED-based luminaires, each of which has a controller that can process information locally and communicate with nearby controllers. The illuminance requirements and energy costs for the lighting system are expressed as a linear programming problem. This optimization problem is solved, in a distributed manner, across the network of controllers using local communication amongst the controllers. State-of-the-art distributed optimization methods, based on accelerated first-order methods, are applied to parallelize the computational tasks among multiple controllers. Important practical aspects such as the rate of convergence, computational complexity, and communication requirements are investigated via simulations