转换观测卡尔曼滤波算法中转换统计量的一种改进计算方法(英文)

为了改进转换观测卡尔曼滤波算法的性能,提出了一种计算转换观测误差统计量(均值和协方差)的改进方法。有两组信息可用于计算这些统计量,即观测值和滤波器的一步预测值。首先,推导了滤波器球坐标状态一步预测误差的均值和协方差。然后,利用滤波器球坐标状态一步预测值更精确地计算了转换观测误差的统计量。最后,将改进的转换观测误差统计量用于目标跟踪问题中的转换观测卡尔曼滤波算法。仿真结果表明,本文提出的方法能够改善转换观测卡尔曼滤波算法的收敛性和估计精度,这种改善在观测噪声较大时尤为显著。

改进路标定位传感模型在机器人定位中的应用

分析了移动机器人基于路标定位的贝叶斯滤波理论中的传感模型,传统传感模型利用周期末检测到的观测值进行更新修正。若在一个周期内不同位置检测到多个不同角度观测值,传统的传感模型默认为这些角度观测值为周期末静止时所得到,提出了通过转化算法将这些角度观测值转换到更新周期末,使得周期末角度观测值得到的位置观测更精确。计算机实验数据表明提出的传感模型观测值转换算法与现有传统算法传感模型相比,更新位置具有更高的精度。

The Least Order Observer Based Sensorless Control of PMSG in Direct-Drive Wind Turbines

Permanent magnet synchronous Generator （PMSG） based direct-drive wind energy conversion system （WECS） has been attracting wide attentions. For the special application, sensorless control for PMSG is desired. By widely studying the previous contributes, a novel estimator based on back-EMF is proposed. The estimator is composed of back-EMF observer and a phase-lock-loop （PLL） control to get the rotor-flux speed and position. The estimator not only can be used for interior and surface permanent magnet synchronous generators, but also has a compact and symmetrical structure, which makes it be beneficial for implementation. Compared with previous strategies, the EMF observer is independent of the PLL control, which would simplify the observer design. Meanwhile, the proposed estimator is less sensitive to parameter variations. Based on mathematic models of PMSG, the proposed estimator was analyzed in detail, and the realizing process was also presented. To validate the proposed estimator, the important experiment results are reported.