Fuzzy-Neural Automatic Daylight Control System

Abstract

The paper presents the design and the tuning of a CMAC controller (Cerebellar Model Articulation Controller) implemented in an automatic daylight control application. After the tuning process of the controller, the authors studied the behavior of the automatic lighting control system (ALCS) in the presence of luminance disturbances. The luminance disturbances were produced by the authors in night conditions and day conditions as well. During the night conditions, the luminance disturbances were produced by turning on and off a halogen desk lamp. During the day conditions the luminance disturbances were produced in two ways: by daylight contributions changes achieved by covering and uncovering a part of the office window and by turning on and off a halogen desk lamp. During the day conditions the luminance disturbances, produced by turning on and off the halogen lamp, have a smaller amplitude than those produced during the night conditions. The luminance disturbance during the night conditions was a helpful tool to select the proper values of the learning rate for CMAC controller. The luminance disturbances during the day conditions were a helpful tool to demonstrate the right setting of the CMAC controller.

Keywords

• daylight control system
• CMAC structure
• basis functions
• B-spline
• learning rate