【摘要】 针对多传感器观测盲区两侧边缘处出现误差较大现象,提出基于阈值的盲区两侧附近异常数据的检测方法。对原有数据沿x轴、y轴方向进行正交分解,分别选取检测这两个方向分量上异常数据的阈值。在此基础上,对原有数据进行筛选,然后通过适当的填补方法,依次对这两个方向上剔除掉的数据进行填补,获得完整的数据集,以替代原有数据。仿真结果表明,采用新算法可以明显提高目标在盲区边缘附近的定位精度。更多还原

【Abstract】 Aiming at the phenomenon that the error becomes larger when the target occurs on either side of the blind spots of the multi-sensor,we have proposed a method detecting the outlier near either side of the blind spots based on the threshold. Along the x axis and y axis,the original data was divided into two groups based on orthogonal decomposition. Then two thresholds were selected to detect the outliers separately. On the basis of this analysis,the original data was discriminated. And then the proper imputation methods were selected to deal with the missing data and get the complete data set. The original data were replaced with the new complete data. Sim-ulation results show that the accuracy of the target positioning near the blind spots can be obviously improved by using this proposed algorithm.更多还原