Application Of Neural Network In Target Detection And Tracking

With the continuous development of electronic technology,the influence of radar on military and civilian use is growing.Radar target detection and tracking is a key issue in radar signal processing.How to improve radar target detection and tracking performance has always been a hot issue.In this paper,the neural network is applied to the weak target detection in the sea clutter background and target tracking to improve its accuracy.The main work is as follows:1.The structure of RBF neural network is mainly determined by experience or trial method,and once the structure is determined,it will not be adjusted,and the defect of adaptive ability is lacking.Firstly,based on adaptive time-varying weight and local search operator,particle swarm optimization is improved.PSO)algorithm,then the parameters of RBF neural network(center value,width,connection weight)as the spatial position of the particle,the spatial dimension of the particle is mapped to the number of neurons in the hidden layer,and the PSO-RBF neural network is constructed,which solves The problem that the RBF structure does not match the parameters greatly enhances the adaptive ability.2.Using the chaotic characteristics of sea clutter,phase space reconstruction of sea clutter time series,combined with PSO-RBF neural network to construct PSO-RBF network target detection model and learning algorithm,and the measured sea clutter data The weak target detection proves its effectiveness and improves the detection accuracy of weak targets in the sea clutter background.3.Aiming at the BP network's choice of BP network learning rate,which is not objective and easy to fall into the local minimum value,the method of increasing the momentum term is proposed to optimize the BP network.Then,the parallel Unscented Kalman Filter is combined with the BP network to construct the tracking model.The target is tracked to improve the tracking accuracy.