针对火炮交流伺服系统中存在的非线性动态摩擦,为了提高火炮伺服系统的跟踪精度,提出了一种基于粒子群的LuGre摩擦参数两步辨识方法。利用Stribeck曲线辨识了4个静态参数,利用滞滑极限辨识了2个动态参数,完成了摩擦力矩的动态补偿。实验结果表明了该算法的合理性,火炮伺服系统的正弦跟踪最大误差为1.25 mrad,精度完全达到了军标要求。 In order to improve the tracking accuracy of artillery servo system, a two-step method of LuGre friction parameter identification based on particle swarm optimization is proposed for the nonlinear dynamic friction existing in artillery AC servo system. Four static parameters were identified by using Stribeck curve. Two dynamic parameters were identified by using the slippage limit, and the dynamic compensation of friction torque was completed. Experimental results show the rationality of this algorithm. The maximum error of sinusoidal tracking of artillery servo system is 1.25 mrad, and the precision meets the requirements of military standard completely.