RBF Neural Network-Based Frequency Band Prediction for Future Frequency Hopping Communications

On the basis of chaotic features frequency hopping signal, band prediction for signal can enhance interference effect greatly. However, poor accuracy often limits its development in military field. Therefore, sake enhancing this paper studies radial function (RBF) neural network model based on gradient descent method and improved particle swarm optimization algorithm, respectively. The former uses a step-by-step algorithm to optimize center value weight so that find most suitable initial state. Then, clustering selection is employed central value. In addition, it optimizes by using optimal learning rate. latter structure RBF through combination subtractive (PSO) algorithm. Simulation results demonstrate performs better terms accuracy, but time efficiency lower, while PSO-RBF has efficiency.