基于量子粒子群优化的动态标定辨识方法

Dynamic calibration and identification method based on quantum-behaved particle swarm optimization

针对传统辨识方法对非线性系统辨识效果不理想的情况,提出将量子粒子群优化(QPSO)算法引入到力传感器的动态标定辨识中来。搭建了基于垂直正弦力加载的力传感器动态标定装置优化,该系统使用正弦运动机构作为激励装置,动态力由安装在正弦机构上的质量块产生。为验证QPSO算法进行系统辨识的可行性,进行了两组对比实验。结果显示:相比于递推最小二乘(RLS)法,QPSO算法的辨识精度较高,适用于非线性系统的参数辨识。

In order to improve unsatisfactory effect of traditional identification methods for nonlinear systems, quantum-behaved particle swarm optimization（QPSO） algorithm is introduced to dynamic calibration identification of force transducers. A dynamic calibration facility with vertical sine force loading is built up. A sine mechanism is employed to generate sinusoidal motion in the facility, and dynamic force can be obtained based on determination of inertia force as a mass mounted on sine mechanism. To verify feasibility of QPSO for system identification, two groups of experiments are carried out. The results indicate that the identification precision of QPSO is superior to recursive least squares method, and it is feasible for nonlinear system identification.