基于UT变换的传感器航迹数据无偏转换

Transform of Sensor Track Data Using Unscented Transformation

在目标跟踪过程中,通常采用笛卡尔坐标系进行航迹输出,这涉及到不同坐标系下航迹数据转换问题。传统变换方法计算量比较大且精度一般。根据UT变换的基本思想,通过选择有限个近似高斯分布离散点,并施以非线性变换,实现了笛卡尔坐标航迹数据的无偏转换。与蒙特卡洛方法的对比分析表明,该方法计算量小,数学推导简单,运算时间短,是一种比较有效的航迹数据转换方法。

Target tracking algorithms typically produce tracks in Cartesian coordinates, but the target range, bearing, elevation, and other estimations are required to meet fusion specifications. An algorithm based on the unscented transformation（UT） is presented to compute such estimates and their error standard deviations. Using the UT approach, the closed-form derivations are unnecessary and computational requirements are an order of the magnitude less than those of the Monte Carlo method. Furthermore, compared with the Monte Carlo method, the UT method does not require a pseudo-random number generator.