一种多传感器配准与目标跟踪算法研究

传统的多传感器误差配准技术多基于球极投影,没有考虑地球地形的影响,当传感器之间距离较远时将失去实际意义,无法对目标进行有效的跟踪;而现有的跟踪方法大多没有考虑传感器系统误差对跟踪精度的影响。基于地心坐标系,提出了一种Unscented卡尔曼配准与目标跟踪算法,充分考虑地球形状的影响,在跟踪目标的同时实现传感器配准。首先给出传感器数据配准几何坐标转换算法,详细推导了误差配准算法;接着建立目标的动态方程,将目标运动模型和传感器配准误差模型组合在同一个状态方程中,然后利用UKF进行估计。最后的Monte-Carlo仿真结果表明,该方法能同时有效地估计目标运动状态和传感器配准误差,为远距离的传感器配准与目标跟踪提供了一种新的解决方法,具有较大的工程应用价值。

An algorithm for multi-sensor registration and target tracking based on UKF

Traditional registration algorithms are all based on the stereographic projection without considering the influence of geometry shape of the earth.Therefore it will be not operable and can't track the target effectively if the distance between sensors is far away.And most of the current tracking methods do not take consideration of the influence of sensor errors.We present here a registration and target tracking algorithm with Unscented Kalman Filter(UKF) based on Earth-Centered Earth-Fixed(ECEF) Coordinate System.This algorithm not only considers the geometry of the earth,but can register sensors while tracking.The algorithm of coordinate conversion of sensor registration is given,and the align algorithm is deduced in detail.Then,a dynamic equation is built up for the target combining target movement model and sensor registration error model in one state equation,and UKF is used to estimate both the registration errors and system states simultaneously.Finally,Monte-Carlo simulation results are used to show the effectiveness of the proposed sensor registration algorithm.The algorithm,which has supplied a new method for registration of sensors apart far-away one another,will play an important role in engineering applications.