An artificial intelligence approach to the prediction of natural lighting levels

Controlling artificial lights within buildings to act solely as a supplement to available daylighting requires continuous knowledge of natural lighting levels within a space. Although this information is readily obtained by measurement whilst lights are extinguished, once illuminated the determination of the underlying natural light level is not so straightforward. This paper describes the use of a Genetic Algorithm (GA) at the heart of a self-commissioning, adaptive algorithm capable of the real-time prediction of natural light levels at chosen points within a room using external measurements of vertical plane illuminance. The algorithm is extremely compact and efficient and readily implemented on a microprocessor. As such, it is suggested that it could form the basis of a robust and practical lighting controller.