Intelligent Illuminance Fusion Algorithm Based on NN-DS

To improve the accuracy of the switching threshold in Intelligent Lighting Control System (ILCS), an Intelligent Illuminance Fusion Algorithm (IIFA) based on the Neural Network (NN) and the Evidence Theory (Dempster-Shafer, D-S) research in this paper. Firstly, this paper proposes an intelligent illuminance fusion model based on the NN and D-S, which respectively using BP NN and RBF NN for feature layer fusion. The output is used in the Basic Probability Assignment (BPA) of the initial evidence. Secondly, in the strategy layer fusion, the combination result of high conflict evidence is often contrary to common sense, so the probability of conflicting evidence is weighted according to the average support degree of each proposition. Finally, through the sensor to collect the sunset and sunrise illumination information, and demonstrate the propose method. Experiments the average accuracy of the fusion result has improved 73.48%. The results indicate the proposed IIFA has higher accuracy and robustness.