To improve and optimize the bandwidth utilization for multi-service packet transporting system, a kind of Dynamic Full Bandwidth Utilized （DFBU） allocation algorithm allowing a single link to use far beyond its fair...To improve and optimize the bandwidth utilization for multi-service packet transporting system, a kind of Dynamic Full Bandwidth Utilized （DFBU） allocation algorithm allowing a single link to use far beyond its fair share bandwidth is presented. Three important parameters as the bound on max and minimum bandwidth, the maximum packet delay and the minimum bandwidth utilization are discussed and analyzed. Results of experiments show that the DFBU-algorithm is capable of making a single link in the system use all the spare bandwidth （up to full-bandwidth） while the performance of fairness and QoS requirement is still guaranteed.展开更多
The wide-band direction finding is one of hit and difficult task in array signal processing. This paper generalizes narrow-band deterministic maximum likelihood direction finding algorithm to the wideband case, and so...The wide-band direction finding is one of hit and difficult task in array signal processing. This paper generalizes narrow-band deterministic maximum likelihood direction finding algorithm to the wideband case, and so constructions an object function, then utilizes genetic algorithm for nonlinear global optimization. Direction of arrival is estimated without preprocessing of array data and so the algorithm eliminates the effect of pre-estimate on the final estimation. The algorithm is applied on uniform linear array and extensive simulation results prove the efficacy of the algorithm. In the process of simulation, we obtain the relation between estimation error and parameters of genetic algorithm.展开更多
A hybrid feature selection and classification strategy was proposed based on the simulated annealing genetic algonthrn and multiple instance learning （MIL）. The band selection method was proposed from subspace decom...A hybrid feature selection and classification strategy was proposed based on the simulated annealing genetic algonthrn and multiple instance learning （MIL）. The band selection method was proposed from subspace decomposition, which combines the simulated annealing algorithm with the genetic algorithm in choosing different cross-over and mutation probabilities, as well as mutation individuals. Then MIL was combined with image segmentation, clustering and support vector machine algorithms to classify hyperspectral image. The experimental results show that this proposed method can get high classification accuracy of 93.13% at small training samples and the weaknesses of the conventional methods are overcome.展开更多
针对农业温室复杂环境中的超宽带(Ultra wide band,UWB)定位精度受非视距(Non line of sight,NLOS)效应和多路径影响的问题,本文提出了一种融合Chan-Taylor与改进沙猫群优化粒子滤波(Chan-Taylor and improved sand cat swarm intellige...针对农业温室复杂环境中的超宽带(Ultra wide band,UWB)定位精度受非视距(Non line of sight,NLOS)效应和多路径影响的问题,本文提出了一种融合Chan-Taylor与改进沙猫群优化粒子滤波(Chan-Taylor and improved sand cat swarm intelligence optimization particle filter,CT+ISCSO-PF)定位算法。首先,利用Chan-Taylor算法实现对目标初始位置的快速估算,为粒子滤波提供准确初值;随后,引入ISCSO(Improved sand cat swarm optimization particle filter)引导粒子向高似然区域移动,通过三角游走策略提升全局搜索能力,结合Levy飞行机制增强局部收敛效率,从而有效抑制粒子退化问题。本文模拟了3种不同噪声水平的环境。仿真结果表明,CT+ISCSO-PF算法在3种环境下,相比于传统的粒子滤波(Particle filter,PF)、Chan-Taylor与粒子滤波(Chan-Taylor and particle filter,CT+PF)、Chan-Taylor与沙猫群优化粒子滤波(Chan-Taylor and sand cat swarm intelligence optimization particle filter,CT+SCSO-PF)、Chan-Taylor与灰狼优化粒子滤波(Chan-Taylor and grey wolf optimizer particle filter,CT+GWO-PF)均表现出明显优势。进一步以农用履带车辆为载体开展温室环境定位试验,结果显示:在LOS场景下,该算法较PF、CT+PF、CT+SCSO-PF和CT+GWO-PF的均方根误差分别降低27.9%、17.8%、7.8%和10.2%;在NLOS场景下,均方根误差降幅分别达21.4%、15.6%、7.6%和5.2%。展开更多
高光谱信息量巨大,如何选取最佳组合波段构建高精度光谱模型,是植被参数遥感反演模型研究的重要工作基础。该研究将最佳指数与相关系数通过熵权评价值进行融合,提出最佳指数-相关系数法(optimum index factor and correlation coefficie...高光谱信息量巨大,如何选取最佳组合波段构建高精度光谱模型,是植被参数遥感反演模型研究的重要工作基础。该研究将最佳指数与相关系数通过熵权评价值进行融合,提出最佳指数-相关系数法(optimum index factor and correlation coefficient,OIFC)。基于OIFC法选取了小麦叶片叶绿素含量的最佳组合波段,并利用最佳组合波段的高光谱数据建立小麦叶片叶绿素含量预测模型。结果表明:利用OIFC法所提取的小麦叶绿素最佳组合波段是760、1 860、1 970 nm;对比最佳指数法(optimum index factor,OIF)、最大相关系数法(maximum correlation coefficient,MCC)提取波段以及归一化植被指数(normalized difference vegetation index,NDVI)、土壤调和植被指数(soil-adjusted vegetation index,SAVI)所建立的叶片叶绿素含量高光谱模型,基于OIFC法构建的模型预测值与实测值具有显著的线性关系,决定系数达0.827,且均方根误差最小(RMSE=5.44)。可见,基于OIFC法构建的小麦叶绿素含量模型具有更高的精度,该结果验证了利用OIFC法提取高光谱特征波段的可行性,并且能够获得更高建模精度的特征波段。展开更多
文摘To improve and optimize the bandwidth utilization for multi-service packet transporting system, a kind of Dynamic Full Bandwidth Utilized （DFBU） allocation algorithm allowing a single link to use far beyond its fair share bandwidth is presented. Three important parameters as the bound on max and minimum bandwidth, the maximum packet delay and the minimum bandwidth utilization are discussed and analyzed. Results of experiments show that the DFBU-algorithm is capable of making a single link in the system use all the spare bandwidth （up to full-bandwidth） while the performance of fairness and QoS requirement is still guaranteed.
基金This project was supported by the Teaching and Research Award Programfor Outstanding Young Teachersin Higher Educa-tion Institutions of MOE (2001226) .
文摘The wide-band direction finding is one of hit and difficult task in array signal processing. This paper generalizes narrow-band deterministic maximum likelihood direction finding algorithm to the wideband case, and so constructions an object function, then utilizes genetic algorithm for nonlinear global optimization. Direction of arrival is estimated without preprocessing of array data and so the algorithm eliminates the effect of pre-estimate on the final estimation. The algorithm is applied on uniform linear array and extensive simulation results prove the efficacy of the algorithm. In the process of simulation, we obtain the relation between estimation error and parameters of genetic algorithm.
文摘A hybrid feature selection and classification strategy was proposed based on the simulated annealing genetic algonthrn and multiple instance learning （MIL）. The band selection method was proposed from subspace decomposition, which combines the simulated annealing algorithm with the genetic algorithm in choosing different cross-over and mutation probabilities, as well as mutation individuals. Then MIL was combined with image segmentation, clustering and support vector machine algorithms to classify hyperspectral image. The experimental results show that this proposed method can get high classification accuracy of 93.13% at small training samples and the weaknesses of the conventional methods are overcome.
