常规伺服系统根据电机轴系转动进行模型分析，以轴系所在的基座空间作为参照系。稳定平台的被控量以惯性空间作为参照系，因此不适合用常规伺服系统模型来建模。针对稳定平台的多参照系问题，文章采用以惯性空间作为电机轴系转动参照系的多空间分析模型，并将改进粒子群算法应用于该模型。粒子群算法作为一种群智能算法，广泛应用于参数优化。文中通过惯性权重改进和越界改进，利用改进后的粒子群算法进行稳定平台PID参数的优化和整定。通过仿真和硬件实验平台验证,结果表明：在稳定平台多空间分析模型基础之上，采用改进粒子群算法优化后的PID控制器可以使稳定平台有更高的稳定精度、更好的鲁棒性，有效地隔离了外部的震动和干扰。

In the conventional servo system, model analysis according to motor axis and base space is used as a reference. However, when analyzing the stabilized platform, it is not compatible due to in which exists inertial space and base space. In order to solve this problem, the multiple reference frame model was proposed, where direct-current motor model was based on inertial space. On the basis of the multiple reference frame model, an improved Particle Swarm Optimization(PSO) algorithm was also proposed. As a kind of swarm intelligence algorithm, PSO was widely used in parameters optimization. The traditional PSO on inertial weight and slopping-over borders were improved, and then, it was adopted in tuning and optimization of PID parameters. The simulation and experiments results indicate that the improved PSO(IPSO) PID controller can obviously enhance the static precision and effectively isolate the vibration and disturbance of carrier.