双激光雷达的水平风场估计方法

Horizontal wind field estimation method based on dual Lidars

针对民航机场区域内单部激光雷达探测水平风场存在较大误差的问题,提出一种基于支持向量回归的双激光雷达水平风场估计模型。该模型以两部激光雷达重叠扫描区域风速为基础,对非重叠区雷达径向的其他数据点水平风速进行估计。首先,提取重叠区的径向风速、水平风速和距离三个特征,以重叠区域数据点为训练集,在同一维度规范化后设定惩罚因子和核函数参数,用支持向量回归得到初始估计值。然后,以单部激光雷达的径向风速为先验条件,估计出非重叠区相邻径向点水平风速。将估计的结果扩展为新的训练集,依次逐步扩大训练集进而估计出非重叠区的水平风速。最后,通过实测数据分析了该方法逐步估计的误差,分析了风速大小和回波信噪比对该方法估计性能的影响,结果表明该方法估计的风场的均方根误差较单部雷达的均方根误差更小,减小了水平风速误差,扩大了双激光雷达探测水平风场范围,提高了雷达的利用率。

Aiming at the problem of error in single Lidar detection of horizontal wind field in civil aviation airport area, a support vector regression based model for estimating horizontal wind field of double Lidars was proposed. The model was based on the wind speed of two Lidars overlapping scanning regions, and the horizontal wind speed at the intersection points was used to estimate the horizontal wind speed of other data points in the radial direction. Firstly, the three characteristics of radial wind speed,horizontal wind speed and distance in the overlapping area were extracted. The overlapping area data points were used as the training set. After the same dimension was normalized, the penalty factor and kernel function parameters were set, and the initial estimated value was obtained by support vector regression. Then, the radial wind speed of the single Lidar was used as the a priori condition to estimate the horizontal wind speed of the adjacent radial points in the non-overlapping area. Then, the radial wind speed of the single Lidar was used as the a priori condition to estimate the horizontal wind speed of the adjacent radial points in the non-overlapping area. The estimated results were extended to a new training set, and the training set was gradually expanded to estimate the horizontal wind speed in the nonoverlapping area. Finally, the error of the stepwise estimation of the method was analyzed by the measured data. The influence of wind speed and echo signal-to-noise ratio on the estimation performance of the method was analyzed. The results show that the root mean square error of the wind field estimated by the method is better than that of the single radar. The method expands the range of the horizontal wind field detected by the dual Lidars and improves the utilization of the Lidars.