粒子群优化的SVM在雷达目标识别中的应用

Application of support vector machine based on particle swarm optimization in radar target recognition

支持向量机(SVM)在雷达目标高分辨距离像(HRRP)识别中可获得较高的正确识别率和更好的泛化性能,然而其性能很大程度上取决于其参数包括核函数参数σ2和惩罚因子C的合理选择。所以利用粒子群优化算法(PSO)全局搜索能力强的优点来搜寻最优参数,并针对粒子群优化易陷入局部最优的问题,提出一种惯性权重自适应改变的改进方法。通过对雷达目标高分辨率距离像(HRRP)的识别实验发现,利用PSO优化SVM参数的方法克服了传统SVM存在的很难精确找到最优参数的缺点,识别准确率也有很大提高;同时惯性权重自适应改变的方法也有效解决了PSO优化的"早熟"问题,大大缩短参数寻优时间。

Support vector machine（ SVM） can obtain higher recognition rate and better generalization ability in the distance radar target high resolution range profile（ HRRP）,but its performance largely depends on selection of the parameters including kernel function σ2parameter and penalty factor C. So it uses the advantages of global search ability of particle swarm optimization algorithm（ PSO） to search for the optimal parameters,and because PSO easily falls into local optimum,this paper proposes an improved adaptive inertia weight change method. In the radar target recognition experiment,PSO-SVM overcomes the shortcomings of the traditional SVM parameter it is difficult to find the optimal parameters,the radar target recognition accuracy rate is greatly improved; the inertia weight adaptive method also can effectively solve the premature problem of PSO,greatly shortens the optimization time.