Distributed Street lighting Model Predictive Control based on Spatial Points of Interest

The street lighting system, as one of the main widely distributed infrastructures of modern urban cities is technologically outdated and suitable for more advanced, remote lighting control. With several thousands of lamps, a real-time centralized control system is technologically and economically challenging. This paper presents a distributed model predictive control algorithm to adjust the illuminance intensity level, i.e. the dimming profile, at chosen spatial points of interest in a street lighting system. The algorithm is formulated as a trajectory-tracking quadratic problem with a central coordinator to deliver dynamically-calculated references at points of interest based on predicted micro location data of traffic volume, pedestrian density and diversity, and weather conditions. Mathematical model of the ray tracing spatial light propagation to ensure horizontal and vertical lighting quality for various urban traffic participants, as well as the reduction of lighting pollution, while improving the energy efficiency of the whole infrastructure. The controller is designed by respecting the premise of the low-cost implementation to large number of distributed lamps. The simulation results are obtained on a real case study data and show the optimal trade-off achieved between illuminance and energy savings.