Entity

Time filter

Source Type


Qu Y.,Jiangxi Normal University | Li Z.,Jiangxi Manufacturing Technology College | Zhang E.,Nanjing University of Science and Technology
Journal of Information and Computational Science | Year: 2011

The Jumping and Static Interacting Multiple Model (JSIMM) is proposed to solve the problem of nonlinear filtering with unknown continuous system parameter. JSIMM regards the continuous system parameter space as a union of disjoint regions, and each region is assigned to a sub-model respectively. The parallel filtering method is adopted in JSIMM for the purpose of responding to the system parameter changing as quickly as possible. For each sub-model, under the assumption that the parameter belongs to the corresponding region, one Unscented Kalman Filter (UKF) is used to estimate the parameter and the state when the parameter is presumed to be jumping, and another UKF is used to estimate the parameter and the state when the parameter is presumed to be static. The results of the two UKFs will be fused as the estimation output of the sub-model. Experiment results show that JSIMM achieves higher performance when compared with IMM, SIR and UKF in bearings only tracking problem. Copyright © 2011 Binary Information Press. Source


Chen Y.,Nanjing University of Science and Technology | Li Z.,Jiangxi Manufacturing Technology College | Jin Z.,Nanjing University of Science and Technology
Neural Processing Letters | Year: 2013

Based on the classification rule of sparse representation-based classification (SRC) and linear regression classification (LRC), we propose the maximum nearest subspace margin criterion for feature extraction. The proposed method can be seen as a preprocessing step of SRC and LRC. By maximizing the inter-class reconstruction error and minimizing the intra-class reconstruction error simultaneously, the proposed method significantly improves the performances of SRC and LRC. Compared with linear discriminant analysis, the proposed method avoids the small sample size problem and can extract more features. Moreover, we extend LRC to overcome the potential singular problem. The experimental results on the extended Yale B (YALE-B), AR, PolyU finger knuckle print and the CENPARMI handwritten numeral databases demonstrate the effectiveness of the proposed method. © 2012 Springer Science+Business Media New York. Source


Keke H.,Nanjing University of Science and Technology | Zhenzhen L.,Jiangxi Manufacturing Technology College | Zhenmin T.,Nanjing University of Science and Technology
International Journal of Digital Content Technology and its Applications | Year: 2011

A problem which may arise from the difference between true mode and model is taken into account, and an improved Interacting Multiple Model algorithm which is based on model error and iterated extended kalman filter. IEKF is used here for each model to handle non-linear estimation problem. The results of Monte-Carlo simulations show that the new algorithm can achieves higher estimation performance than IMM. Source


Li Z.,Jiangxi Manufacturing Technology College | Cai T.,Jiangxi Manufacturing Technology College
Journal of Computational Information Systems | Year: 2013

Based on ridge regression (RR), we propose a novel method called simplexface approach for face recognition. RR treats each individual as a regular simplex vertex. Then the training samples are mapped to the corresponding vertices as near as possible. In essence, RR only considers the between-class scatter. Thus, the incompactness of the within-class samples may lead to inseparability. Different from RR significantly, the simplexface approach is a complete discriminant analysis method since it considers both betweenclass and within-class scatter. The simplexface approach increases the within-class compactness which further strengthens the separability. Experimental results on the AR face database demonstrate the effectiveness of our method. Copyright © 2013 Binary Information Press. Source

Discover hidden collaborations