基于多项式统计的舰船传感器偏差估计算法

Polynomial and Statistical Algorithm of Bias Estimation for Ship Sensor

现代舰艇配置多部用于探测作战任务目标的传感器,因此必须估计距离、方位和俯仰的探测参数偏差;大部分已有算法需要从传感器获取额外信息,比如滤波增益和关联协方差矩阵;文章提出7阶多项式拟合和假设检验的新算法,使用K-S检验、卡方检验和 t 检验方法统计分析估计传感器系统偏差;通过比较不同传感器的航迹数据,该算法可获得多种传感器的探测精度和偏差,并提供传感器间偏差异常定位;最后,通过仿真数据和无人机测量数据验证本文所提算法的有效性。

Modern warships possess several sensors to detect targets for combat mission, so it is necessary to estimate bias of range, azimuth, elevation and velocity parameters. Most previous bias-estimation algorithm requires additional information such as filter gain and associated covariance matrices from sensors, while this novel algorithm uses statistical analysis for bias estimation of these parameters for ship sensors, with 7-level minimum-variance polynomial fitting and null hypothesis testing. By statistically comparing track data from different sensors, this algorithm locates sensor abnormality of measurement accuracy for sensors and the bias between different these sensors. This algorithm is verified by sensor measurement of unmanned aerial vehicle (UAV).