北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2012, Vol. 35 ›› Issue (2): 90-93.doi: 10.13190/jbupt.201202.90.wuch

• 研究报告 • 上一篇    下一篇

GA-LSSVM的认知无线电离线学习

伍春1,2,李颖3,易克初1   

  1. 1. 西安电子科技大学 综合业务网理论及关键技术国家重点实验室, 西安 710071; 2. 西南科技大学 国防科技学院, 四川 绵阳 621000; 3. 中国电子系统工程公司研究所, 北京 100141
  • 收稿日期:2011-07-29 修回日期:2011-10-17 出版日期:2012-04-28 发布日期:2012-01-05
  • 通讯作者: 伍春 E-mail:soldier_wu@163.com
  • 作者简介:伍春(1978-),男,博士生,E-mail:soldier_wu@163.com 易克初(1943-),男,博士生导师
  • 基金资助:

    国家重点基础研究发展计划项目(2009CB320403);国家科技重大专项项目(2009ZX03007-004);国家自然科学基金项目(61072138);西安电子科技大学ISN实验室开放课题(ISN10-09)

GA-LSSVM Offline Learning in Cognitive Radios

WU Chun1,2,LI Ying3,YI Ke-chu1   

  1. 1. State Key Laboratory of Integrated Service Networks, Xidian University, Xi-an 710071, China;2. School of National Defense Technology, Southwest University of Science and Technology, Sichuan Mianyang 621000, China;3. Institute of China Electronic System Engineering Corporation, Beijing 100141, China
  • Received:2011-07-29 Revised:2011-10-17 Online:2012-04-28 Published:2012-01-05
  • Contact: Chun WU E-mail:soldier_wu@163.com

摘要:

针对认知无线电(CR)的智能学习研究需求,提出了一种基于最小二乘支持向量机(LSSVM)的CR学习、决策方法. 通过遗传算法(GA)进行超参数搜索,建立适当的基于LSSVM的CR学习模型,利用历史案例进行LSSVM离线训练学习,获得知识来指导未来的决策. 仿真结果表明,GALSSVM能在较少进化代数内搜索出合适的超参数,并且LSSVM学习决策方法能有效提高CR系统的性能.

关键词: 认知无线电, 智能学习, 最小二乘支持向量机, 遗传算法

Abstract:

A cognitive radio (CR) learning and decision making method based on least squares support vector machine (LSSVM) is proposed by the demand of CR intelligent learning research. Genetic algorithm is adopted to search the hyper-parameters of LSSVM, and the CR learning model based on LSSVM is established. Through LSSVM offline learning on historical instances, the CR system gets the knowledge and uses it to guide future decisions. Simulation shows that GA-LSSVM algorithm can find out appropriate hyper-parameters within small generations and the LSSVM learning and decision making method can improve CR system’s performances effectively.

Key words: cognitive radio, intelligent learning, least squares support vector machine, genetic algorithm

中图分类号: