关键词: 人工免疫系统, FCM聚类, RBF神经网络, 信息融合, 交通

Abstract: In the research of radial basis function neural network applied to traffic information fusion, a fuzzy C-means (FCM) clustering algorithm has been used to determine the radial center of a hidden layer network. This method generally randomly initializes the clustering centers, and the training process is easy to get into a local minimum. With the clonal selection theory in the artificial immune systems, a new algorithm used for calculating the initial clustering centers was presented. By integrating a modified artificial immune algorithm with FCM clustering algorithm, RBF structural parameters were optimized. This algorithm can overcome the shortcomings of getting into local optimization from traditional algorithms, increase the convergent rate, and obtain better fusion effects. Experiments verify the efficiency and practicability of this algorithm.

Key words: artificial immune system, FCM clustering, RBF neural network, information fusion, traffic