Target Recognition Research Based On PSO-SFLA And UWB

In recent years,Ultra-Wideband(UWB)wireless communication technology has been widely used in the target detection and recognition under foliage environment because of its strong anti-interference,anti-multipath ability and UWB has excellent penetration ability to leaves.Low detection rate and high false alarm rate are very common in the detection and recognition of target under foliage environment.Ultra-Wideband signal has a great advantage in this aspect,but the high efficiency and high accuracy rate are still the goal pursued by researchers.Aiming at the problem of efficiency of target recognition,this paper uses the time domain feature extraction of target detection,which decreases the information redundancy,and the target recognition is more effective;we compare the performance of several common classifiers,and select the support vector machine as the classification model to improve the efficiency of target recognition.In order to ensure’ the accuracy of the recognition rate,it is necessary to guarantee a certain number of time domain characteristic parameters,but it also makes the recognition speed is slow.So the dimension compression is achieved by using principal component analysis(PCA).The experimental results show that the proposed method can greatly improve the recognition efficiency.In this paper,aiming at the problem of accuracy of target recognition,the target recognition is realized based on the support vector machine,the recognition performance of this classifier is greatly affected by the parameters.The existing parameter optimization algorithms have advantages and disadvantages in the aspects of convergence speed and global optimal performance.This paper uses particle swarm algorithm(PSO)with excellent convergence rate to optimize the shuffled frog leaping algorithm(SFLA)with strong ability of global search.The improved algorithm(PSO-SFLA)has fast convergence speed and global optimal ability,then PSO-SFLA is used to optimize the parameters of SVM.Experimental results show that the PSO-SFLA has more excellent performance in improving the accuracy of target recognition compared with SFLA and PSO.