为解决路侧毫米波雷达数据分布不均匀、噪点较多,导致雷达在多目标跟踪方面准确性下降的问题,本文在密度峰值聚类(density peaks clustering,DPC)算法的基础上,提出一种适用于毫米波雷达数据改进的DPC算法。该方法结合匈牙利匹配算法和...为解决路侧毫米波雷达数据分布不均匀、噪点较多,导致雷达在多目标跟踪方面准确性下降的问题,本文在密度峰值聚类(density peaks clustering,DPC)算法的基础上,提出一种适用于毫米波雷达数据改进的DPC算法。该方法结合匈牙利匹配算法和无迹卡尔曼滤波算法(unscented Kalman filter,UKF),实现77 GHz毫米波雷达在不同车流密度路面上的跟踪验证。实验结果表明:改进后的DPC算法能够在大幅度提高毫米波雷达数据聚类精度的同时,有效提高多目标跟踪算法的跟踪精度。展开更多
In recent years,polarization remote sensing has garnered increasing attention,particularly within the realm of meteorology.To accurately simulate polarization information,the vector discrete-ordinate radiative transfe...In recent years,polarization remote sensing has garnered increasing attention,particularly within the realm of meteorology.To accurately simulate polarization information,the vector discrete-ordinate radiative transfer(VDISORT)model developed earlier by the community is further enhanced to an advanced version(referred to as A-VDISORT)through an improved ocean surface reflection.The Fresnel reflection matrix,which includes wind-generated roughness and shading effects,is served by an ocean bidirectional reflection distribution function(BRDF).The simulation from AVDISORT is compared with SCIATRAN for a Rayleigh scattering atmosphere,and the influence of water-leaving radiance is analyzed by the PSTAR(Polarized System for Transfer of Atmospheric Radiation) model.For GaoFen-5 Directional Polarimetric Camera(DPC) observations with polarization and multi-angle information,clear-sky pixel recognition over the ocean is first carried out.The DPC reflectance of clear conditions is normalized and compared with the observations.It is shown that A-VDISORT has a high simulation accuracy with a bias of –0.0053.The difference between simulation and observation exhibits a standard normal probability distribution function.展开更多
文章针对生产过程中质量数据分布类型未知引起的传统质量控制图异常检测精度低的问题,提出结合支持向量数据描述(support vector data description,SVDD)和密度峰值聚类(density peaks clustering,DPC)的制造过程异常检测方法。采用DPC...文章针对生产过程中质量数据分布类型未知引起的传统质量控制图异常检测精度低的问题,提出结合支持向量数据描述(support vector data description,SVDD)和密度峰值聚类(density peaks clustering,DPC)的制造过程异常检测方法。采用DPC算法对质量特征数据进行聚类分析,将聚类结果作为模型输入训练得到各类超球体中心和决策边界;以此建立基于内核距离的DPC控制图,实现对生产过程质量波动的实时监控;最后将该控制图应用到再制造曲轴生产过程监控中。结果表明,该文提出的DPC控制图可以有效监测再制造曲轴生产过程质量异常波动,验证了该检测方法的可行性和有效性。展开更多
针对间歇采样转发干扰产生的假目标和目标高速运动产生的多普勒频移导致雷达脉压性能急剧下降的问题,提出一种高多普勒容限的线性调频离散相位编码(linear frequency modulation-discrete phase coding,LFM-DPC)复合调制相干波形集设计...针对间歇采样转发干扰产生的假目标和目标高速运动产生的多普勒频移导致雷达脉压性能急剧下降的问题,提出一种高多普勒容限的线性调频离散相位编码(linear frequency modulation-discrete phase coding,LFM-DPC)复合调制相干波形集设计方法。在一定多普勒频移范围内,以最小化未转发信号自模糊函数旁瓣能量以及未转发信号与转发信号互模糊函数能量建立优化模型,并设计一种基于KKT(Karush-Kuhn-Tucker)最优性条件的迭代算法对模型求解。仿真实验表明,相比于遗传算法和单一调制的LFM和DPC信号,基于KKT最优性条件的交替迭代优化算法优化的LFM-DPC波形集有更好的抗间歇采样转发干扰性能。展开更多
基金supported by the National Natural Science Foundation of China(Grant No.U2142212)National Key R&D Program of China(Grant No.2019QZKK,Qinghai Tibet KeKao)the National Natural Science Foundation of China(Grant No.U2242211)。
文摘In recent years,polarization remote sensing has garnered increasing attention,particularly within the realm of meteorology.To accurately simulate polarization information,the vector discrete-ordinate radiative transfer(VDISORT)model developed earlier by the community is further enhanced to an advanced version(referred to as A-VDISORT)through an improved ocean surface reflection.The Fresnel reflection matrix,which includes wind-generated roughness and shading effects,is served by an ocean bidirectional reflection distribution function(BRDF).The simulation from AVDISORT is compared with SCIATRAN for a Rayleigh scattering atmosphere,and the influence of water-leaving radiance is analyzed by the PSTAR(Polarized System for Transfer of Atmospheric Radiation) model.For GaoFen-5 Directional Polarimetric Camera(DPC) observations with polarization and multi-angle information,clear-sky pixel recognition over the ocean is first carried out.The DPC reflectance of clear conditions is normalized and compared with the observations.It is shown that A-VDISORT has a high simulation accuracy with a bias of –0.0053.The difference between simulation and observation exhibits a standard normal probability distribution function.
文摘文章针对生产过程中质量数据分布类型未知引起的传统质量控制图异常检测精度低的问题,提出结合支持向量数据描述(support vector data description,SVDD)和密度峰值聚类(density peaks clustering,DPC)的制造过程异常检测方法。采用DPC算法对质量特征数据进行聚类分析,将聚类结果作为模型输入训练得到各类超球体中心和决策边界;以此建立基于内核距离的DPC控制图,实现对生产过程质量波动的实时监控;最后将该控制图应用到再制造曲轴生产过程监控中。结果表明,该文提出的DPC控制图可以有效监测再制造曲轴生产过程质量异常波动,验证了该检测方法的可行性和有效性。
文摘针对间歇采样转发干扰产生的假目标和目标高速运动产生的多普勒频移导致雷达脉压性能急剧下降的问题,提出一种高多普勒容限的线性调频离散相位编码(linear frequency modulation-discrete phase coding,LFM-DPC)复合调制相干波形集设计方法。在一定多普勒频移范围内,以最小化未转发信号自模糊函数旁瓣能量以及未转发信号与转发信号互模糊函数能量建立优化模型,并设计一种基于KKT(Karush-Kuhn-Tucker)最优性条件的迭代算法对模型求解。仿真实验表明,相比于遗传算法和单一调制的LFM和DPC信号,基于KKT最优性条件的交替迭代优化算法优化的LFM-DPC波形集有更好的抗间歇采样转发干扰性能。
