In some object tracking systems, the moving object future position is an area （i.e., target area）. It is a successful estimation strategy if the predicted points fall in the target area. If the object makes a sudden...In some object tracking systems, the moving object future position is an area （i.e., target area）. It is a successful estimation strategy if the predicted points fall in the target area. If the object makes a sudden maneuvering, the prediction may get out of the target area easily which may make the tracking system lose the object. The aim is to investigate the admissible maximum object maneuvering intensity, which is characterized as model noise variance, for such kind of tracking system. Firstly, the concept of stochastic passage characteristics over the boundary of target area and their relationship with prediction error variance are described. Secondly, the consistency among the indices of regional pole, prediction error variance and stochastic passage characteristics is analyzed. Thirdly, the multi-indices constraints are characterized by a set of bi-linear matrix inequalities （BMIs）. Then, the admissible maximum model noise variance and the satisfactory estimation strategy are presented by iteratively solving linear matrix inequalities （LMIs） to approximate BMIs. Finally, a numerical example is proposed to demonstrate the obtained resuits.展开更多
Aiming at some weapon systems with shooting domain,the stochastic passage characteristics of the barrel were studied.On the basis of the exact definition of the stochastic passage characteristics,its opportunity-await...Aiming at some weapon systems with shooting domain,the stochastic passage characteristics of the barrel were studied.On the basis of the exact definition of the stochastic passage characteristics,its opportunity-awaiting time,residence time and stochastic passage period were given by using the transition probability matrix,and they all obeyed the geometry distributions.Their means and variances were also derived,and the relations between the time indexes and the structure and parameters of weapon control system were established.Finally,the creditability of the conclusions was verified by the test data of weapon system in proving ground.展开更多
多视图聚类方法随着数据获取途径日益多样化成为研究热点,但大多数聚类方法低估了噪声和数据多结构互补性信息对聚类结果的影响,并且忽略了聚类结果对低秩张量优化过程的反向引导作用。为解决这些问题,提出了基于结构化张量学习的多视...多视图聚类方法随着数据获取途径日益多样化成为研究热点,但大多数聚类方法低估了噪声和数据多结构互补性信息对聚类结果的影响,并且忽略了聚类结果对低秩张量优化过程的反向引导作用。为解决这些问题,提出了基于结构化张量学习的多视图聚类(multi-view clustering based on structured tensor learning,MCSTL)。首先,对初始表示张量进行再次去噪使其更具准确性和鲁棒性;同时,互补地学习局部结构、全局结构和各视图间的高阶相关性,提高表示张量与原始数据本质簇结构的一致性;然后,从跨视图信息融合的亲和矩阵中学习到统一的特征矩阵,利用其隐含的聚类结构信息反向引导表示张量的优化过程;最后,对特征矩阵施加了正交约束,使其提供数据的软标签信息,并对模型进行直接聚类解释。实验表明,MCSTL在6种聚类评价指标上均表现优异,30个指标数据中有27个达到最优,从而充分验证了MCSTL的有效性和优越性。展开更多
基金supported by the Science and Technology Development Fund of Nanjing University of Science and Technology(NUST)(XKF09020)NUST Research Fund(2010GJPY067,2010ZYTS050)the National Natural Science Foundation of China(60804019)
文摘In some object tracking systems, the moving object future position is an area （i.e., target area）. It is a successful estimation strategy if the predicted points fall in the target area. If the object makes a sudden maneuvering, the prediction may get out of the target area easily which may make the tracking system lose the object. The aim is to investigate the admissible maximum object maneuvering intensity, which is characterized as model noise variance, for such kind of tracking system. Firstly, the concept of stochastic passage characteristics over the boundary of target area and their relationship with prediction error variance are described. Secondly, the consistency among the indices of regional pole, prediction error variance and stochastic passage characteristics is analyzed. Thirdly, the multi-indices constraints are characterized by a set of bi-linear matrix inequalities （BMIs）. Then, the admissible maximum model noise variance and the satisfactory estimation strategy are presented by iteratively solving linear matrix inequalities （LMIs） to approximate BMIs. Finally, a numerical example is proposed to demonstrate the obtained resuits.
基金Sponsored by National Defense Fundation of China(9140C300602080C30)NUST Research Fundation of China(2010ZYTS050)
文摘Aiming at some weapon systems with shooting domain,the stochastic passage characteristics of the barrel were studied.On the basis of the exact definition of the stochastic passage characteristics,its opportunity-awaiting time,residence time and stochastic passage period were given by using the transition probability matrix,and they all obeyed the geometry distributions.Their means and variances were also derived,and the relations between the time indexes and the structure and parameters of weapon control system were established.Finally,the creditability of the conclusions was verified by the test data of weapon system in proving ground.
文摘电站辅机设备健康状态评估与故障预警对新型电力系统火电机组的安全运行具有重要意义。以某超临界660 MW火电机组送风机为研究对象,提出了一种基于多重特征参数的送风机故障模型动态记忆矩阵构建方法,该方法可在确保计算结果精度的同时有效提升模型计算速度。同时引入权重系数改进多元状态估计(multivariate state estimation technique,MSET)算法,提出了一种权重系数计算方法;采用总体相似度和参数相似度指标进行故障预警和定位,构建了基于动态记忆矩阵和加权MSET算法的送风机故障预警模型。运用该模型对送风机故障进行仿真,仿真结果表明:加权MSET算法不仅能够有效提高故障工况下异常参数的预测精度,还能降低异常参数对正常参数预测结果的影响,进而在实现送风机故障提前预警的同时准确定位出故障点参数。
文摘多视图聚类方法随着数据获取途径日益多样化成为研究热点,但大多数聚类方法低估了噪声和数据多结构互补性信息对聚类结果的影响,并且忽略了聚类结果对低秩张量优化过程的反向引导作用。为解决这些问题,提出了基于结构化张量学习的多视图聚类(multi-view clustering based on structured tensor learning,MCSTL)。首先,对初始表示张量进行再次去噪使其更具准确性和鲁棒性;同时,互补地学习局部结构、全局结构和各视图间的高阶相关性,提高表示张量与原始数据本质簇结构的一致性;然后,从跨视图信息融合的亲和矩阵中学习到统一的特征矩阵,利用其隐含的聚类结构信息反向引导表示张量的优化过程;最后,对特征矩阵施加了正交约束,使其提供数据的软标签信息,并对模型进行直接聚类解释。实验表明,MCSTL在6种聚类评价指标上均表现优异,30个指标数据中有27个达到最优,从而充分验证了MCSTL的有效性和优越性。
