Font Size: a A A

Study On Support Vector Machine And Its Application In MIMO-CDMA System Multiuser Detection

Posted on:2012-01-05Degree:MasterType:Thesis
Country:ChinaCandidate:P F GaoFull Text:PDF
GTID:2218330362452287Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
The Support Vector Machine (SVM) is a learning method based on statistical theory, SVM is better at non-linear processing and generalization than the other general learning machine, meanwhile, it has better global convergence. Therefore, the SVM can solve the problems of multi-user detection in MIMO-CDMA system First, this paper describes the key theoretical problems , which is used in the derivation of SVM, for example,the dual theory, convex programming,the use of kernel functions and so on. Then, it focus on improved algorithms on support vector machine,especially the new branch of SVM ----relevance vector machine. Finally, MIMO-CDMA system multi-user detection based on SVM is studied.This paper's main work is as follows:(1) Study on SVM and its improved algorithm. There are a variety of improved SVM algorithms, whose advantages and disadvantages exist simultaneously. On the basis of previous studies, quantum particle swarm optimization (QPSO) is introduced to solve quadratic programming problems in SVM. In addition, derivation of support vector regression machine principle is not consistent, the new view is proposed in this paper, which can give concise and easily understandable derivation.(2) Study on Relevance vector machine. Relevance vector machine was proposed recently, which needs to be further improved in many ways, particularly, Matrix inversion and iterative Solution of hyper parameters in the algorithm. These limit the scope of application of the relevant vector machine. After generalized relevance vector machine is in-depth studied, work is done in the above two areas described .(3) Study on MIMO-CDMA multi-user detection. On the basis of classical model of multi-user detection method, RVM is introduced into the received signal detection to design multi-user detection system. The comparisons of simulations demonstrate that the system has good performances of anti-MAI, and it can achieve the desired effect.
Keywords/Search Tags:support vector machines (SVM), kernel function, relevance vector machine (RVM), quantum particle swarm optimization (QPSO), hyperparameter, MIMO-CDMA multi-user detection
PDF Full Text Request
Related items