Rsearch On The Applications Of Neural Network Based On Immune Algorithm

Intelligent signal processing is concerned with simulating biological processing methods to acquire intelligent information processing functions. Artificial neural networks (ANNs) are based on the recognition of human brain neuron networks by modern neurobiology. Through inter-connection of massive neuron collection, ANNs process information in a parallel way which is derived from their self-learning, self-organization and nonlinear dynamics. Artificial immune systems (AIS) are information processing systems based on the metaphor of relevant mechanisms in natural immune systems. In recent years, AIS has developed very quickly, and has become anther active research branch in intelligent signal processing area. In this paper immune optimization algorithm and radial basis function(RBF) neural network are discussed; furthermore, RBF network trained with immune optimization algorithm is applied to the synthetically sorting of radar signals, which achieves very good performance. The main work of the paper can be summarized as follows:1. Immune optimization algorithm is an important branch of AIS. It incorporates the evolutionary mechanism of natural immune system into canonical evolutionary algorithm, and provides new methods for stochastic searching. In the paper many immune optimization algorithms based on different immunological metaphors are introduced, and some of the algorithms are tested with TSP problems, which demonstrate the performance improvement of immune optimization algorithms.2. Radar signal sorting is a basic signal processing technique in radar reconnaissance receivers. Only on the basis of signal sorting can the style and threatening degree of radar be effectively identified. In the paper RBF network based on immune optimization algorithm is applied to synthetically sorting system of radar signals. The effectiveness of the proposed RBF sorting system is demonstrated through experiments under complex radar signal environments.3. OFDM channel estimation is one of the key techniques in wireless communication. A channel estimation method based on a novel RBF network using artificial immune algorithm is proposed in OFDM channel estimation. In this method, the nonlinear parameters of RBF hidden layer are determined by an immune algorithm, which can effectively overcome the immature problem in the evolutionary algorithm. Computer simulations with channel under low SNR and fast fading circumstances, demonstrate that RBF network estimator in this method can rebuild the transform function accurately, and is an effective method to estimate the channel in OFDM system.