基于双重扩展卡尔曼滤波器的共轴跟踪技术研究

On-axis tracking technology based on the dual extended Kalman filter

为了解决光电经纬仪由于机动目标运动模型不准确而引起的跟踪精度下降的问题,采用了单隐层前向神经网络(SLFNs)进行建模,提出了基于状态参数双重扩展卡尔曼滤波估计的共轴跟踪控制技术。仿真与实验结果显示,对83.33°sin0.6t的等效正弦目标的速度估计最大误差为0.070 9(°)/s,跟踪精度为2.42′;对旋转周期为4.5 s的光学动态靶标的跟踪精度达到2.96′以内。由此可见,所建立的模型与机动目标实际模型匹配,双重扩展卡尔曼滤波器(DEKF)能快速跟踪和估计状态参数。与传统控制方法相比,提出的方法具有更高的跟踪能力,能有效提高系统的跟踪精度。

In order to solve the problem of the declining tracking accuracy for O-E theodolite caused by maneuvering target model inaccurate, the single hidden layer feedforward networks（SLFNs） model was built for maneuvering target, then an dual extended Kalman filter（DEKF） on-axis tracking control method was proposed. The key state parameters of the mathematical model were estimated by the dual extended Kalman filter（DEKF）. In the simulation and experiments, the predicted target velocity error is about 0.070 9（ °）/s at the peak of velocity and the tracking accuracy is 2.42 ′ when tracking the equivalent sine object of83.33° sin0.6t, the tracking accuracy reaches 2.96′ when tracking the optical dynamic target with 4.5 s rotation period. Tracking results show that SLFNs mode is matched for maneuvering target, DEKF can fast track and estimate the state parameters. As compared with other control methods, the proposed method has higher accuracy and improves the system tracking precision significantly.