频域和时域船舶动力定位参数辨识算法比较

摘要

针对频域方法和时域方法均被用于船舶动力定位系统建模,但未见特点和适用性对比研究的问题,以船舶动力定位系统参数辨识为对象,进行了频域法R-MISO (reverse multiple input-single output)和时域算法EKF (extended kalmanfilter)的算法实现难易程度、收敛速度、参数依赖性、抗噪能力和适应场合等的比较；基于MATLAB仿真结果显示RMISO算法辨识时间短,不需要设置初始值,适用于在线辨识,而EKF算法去噪能力比较强,需要数据量大,结果依赖初始值,适用于离线辨识,从而为实际船舶动力定位系统参数辨识提供一定的指导.%The methods of frequency domain and time domain are used in the ship dynamic positioning system model,in order to find the comparison of characteristics and applicability,the methods of frequency domain method R-MISO (reverse multiple input-single output) as well as the time domain identification method EKF (extended kalman filter) are taken into consideration in ship dynamic positioning system parameters identification,and carry out a comparative study of difficulty degree,convergence speed,parameter dependence,noise immunity ability and adapt to the comparison of the occasion of the methods.By the MATLAB simulation,the two algorithms are compared,the results show that the R-MISO algorithm is faster,and it doesn't need to provide the initial value,it can be better applied to the online identification,on the other hand the EKF method processes a stronger denoising ability,however it needs a larger number of data as well as the dependence of the initial value,it can be better applied to the offiine identification,to some degree a guidance is provided for the identification process of the ship dynamic positioning system.