【摘要】 针对目前研究的时间配准方法是在目标运动模型已知的情况下进行时间配准,难以保证目标在复杂机动情况下运动模型多变时的时间配准精度。提出了机动目标的交互多模型扩展卡尔曼滤波(IMM-EKF)时间配准算法,该算法将交互多模型中的每个运动模型分别进行扩展卡尔曼滤波输出同时根据滤波过程中得到的残差计算每个模型的概率,根据模型概率和各模型滤波输出得到时间配准周期内最后一个采样点的测量数据,利用该点的状态和模型概率进行外推就得到时间配准周期和传感器采样周期不成整数比时配准时刻的位置。通过仿真结果表明该算法能够有效降低整体的时间配准误差。该算法提高了时间配准的精度,为数据融合提供了良好的基础。更多还原

【Abstract】 Now time registration process is researched at the situation of the target model known. In fact,it is difficult to make ensure accuracy of the time registration when sports model of the maneuvering target are always varied and not known previously. This paper presents an algorithm on IMM extended Kalman filter( IMM-EKF) time registration based on maneuvering target. In the algorithm,each motion model were output by extended kalman filter while residues obtain by the filtering process differential probability to calculate for each model,and use the model probability and output of each model to calculate last sample point state estimation,then use the point of state and probabilistic models to extrapolate to obtain the registration time position when ratio between the period of time registration and the period sensor sampling is not an integer. The simulation results show that the algorithm can effectively reduce the overall time of registration error. The algorithm improves the accuracy of the registration period for data fusion provides a good foundation.更多还原