基于UKF-IMM的双红外机动目标跟踪算法

为了有效解决红外机动目标跟踪精度问题,提出基于UKF的交互式多模型IMM红外机动目标跟踪算法.该方法采用Markov过程描述多个目标模型间的切换,同时导出滤波器输入输出均加权的交互式算法.滤波器采用UKF,避免计算扩展卡尔曼滤波EKF所需的Jacobi矩阵,适用于非线性、非高斯的目标系统模型和观测模型,同时UKF可供多个模型共用,便于软、硬件实现.最后,用双红外探测器对S型机动目标进行仿真实验,给出应用该方法的具体步骤,验证了IMM-UKF的稳定性、有效性和精确性.

Maneuvering target passive tracking with dual infrared observers using IMM algorithm based on UKF

For tackling the accuracy problem of tracking a maneuvering target by infrared sensors,an algorithm of UKF-IMM(unscented Kalman filter-interacting multiple model) is presented.This method use Markov process to describe swiching probability among the models,while weighting means of inputs and outputs of filters.UKF is a good algorithm for nonlinear,noguassian system models and measure models,while commonly used by all models.UKF needn't compute Jacobi matrix of EKF(extended Kalman filter),easier to implement by software or hardware.Finally,simulation results in passive maneuvering target tracking using dual infrared sensors show that the proposed method is more stable,effective,and better accuracy.