Search

Within the learning framework of maximum weighted likelihood (MWL) proposed by Cheung, 2004 and 2005, this paper will develop a batch Rival Penalized Expectation-Maximization (RPEM) algorithm for density mixture clustering provided that all observations are available before the learning process. Compared to the adaptive RPEM algorithm in Cheung, 2004 and 2005, this batch RPEM need not assign the learning rate analogous to the Expectation-Maximization (EM) algorithm (Dempster et al., 1977), but still preserves the capability of automatic model selection. Further, the convergence speed of this batch RPEM is faster than the EM and the adaptive RPEM in general. The experiments show the superior performance of the proposed algorithm on the synthetic data and color image segmentation.

This paper propose a method to deal with the complex nonlinear characteristics, which was caused by the uncertainty of the unknown. As the received signal strength is low in Non line-of-sight environment, the proposed algorithm can be used to process and predict the unknown nodes position by maximum likelihood estimation method. Furthermore, experimental results show that the mobile target tracking error can be less than 1m when we use the method mentioned above, thus meeting the requirements of smart space mobile targeting.

This paper proposes a new fast method for identifying the mixing matrix based on a binary state mixture of Gaussian (MoG) source model. First, a necessary discussion for solving the mixing matrix detection is offered under the multiple dominant circumstance. Second, a density detection method is presented to improve the identification performance. Simulations are given to demonstrate the effectiveness of our proposed approach.

Based on Bernoulli-Gaussian model and the sparseness of source signals, a new algorithm for estimating the mixing matrix is proposed in this paper. It estimates the mixing matrix by searching the cluster points which are found through the density of points in the region. In order to enhance the precision of the algorithm, the cost function is constructed to search the cluster points. The last simulations show the good performance of the proposed algorithm.

For the problem of Underdetermined Blind Source Separation (UBSS) under nonstrictly sparse condition, we propose a new algorithm to estimate mixing matrix. Firstly, a clustering prototype of orthogonal complement space is introduced; Secondly, a fuzzy EVD clustering method which combines fuzzy clustering and Eigenvalue Decomposition (EVD) is presented. Based on these two methods, the algorithm proposed in this paper is robust against noise without losing convergence speed.