基于自适应训练的一种新的软决策融合算法  

New Soft Decision Fusion Arithmetic Based on Adaptive Training

作　　者：刘英坤[1] 冯新喜[1] 党宏刚[1] 潘平俊[1] 

机构地区：[1]空军工程大学电讯工程学院,陕西西安710077

出　　处：《现代防御技术》2008年第3期70-73,共4页Modern Defence Technology

摘　　要：为了提高系统的检测性能,本地检测器应用了软决策算法,提出一种新的结构框图。将本地判决送至自适应训练器来估计本地检测器的性能参数,然后将得到的性能参数反馈至本地检测器来控制本地的检测门限。另外又将本地判决和可信度送至融合中心进行融合判决得到最终的判决,随后给出了迭代算法。最后通过三检测器系统仿真说明了算法的可行性,检测性能比硬判决有所改善。In order to improve detective performance,soft detection is applied in local detectors,a new structure figure is brought forward.The local decisions are communicated to the adaptive training to approximate the performance parameters of the local detectors that are feed back to control the thresholds of local detectors.In addition the local decision and reliability are communicated to the fusion center to make the ultimate decision.Later the iterative arithmetic is presented,at last based on three-detector s...

关 键 词：分布式软决策 自适应训练 聂曼-皮尔逊检测规则 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]