基于支持向量机的传感器管理技术研究  

Study on sensor management based on support vector machine

作　　者：蒋鼎国[1] 张宇林[1] 焦竹青[1] 徐保国[1] 

机构地区：[1]江南大学通信与控制工程学院,江苏无锡214122

出　　处：《传感器与微系统》2008年第11期21-23,共3页Transducer and Microsystem Technologies

基　　金：国家"863"计划资助项目(2006AA10A301)

摘　　要：传感器的优化管理是影响多传感器目标分配问题的重要环节。在分析传感器管理基本要求的基础上,提出了一种基于支持向量机(SVM)的新型目标判定方法。利用SVM算法能够对经验风险和置信范围进行有效控制的优点,以进行多传感器系统的融合预测与可靠性检测,最终实现了在实测数据和可靠的目标预测信息下,对传感器资源的有效管理。仿真实例证明了该传感器管理方法的合理性和有效性。Sensor optimal management is very important to multi-sensor system in target assignment. To solve the problem effectively, the basic requirement of sensor management are analyzed, and a novel target assignment algorithm is proposed based on support vector machine （SVM） , Through the active control to the experience risk and belief range using SVM, the fusion forecast and reliable examination of multi-sensor system is carried on. Finally, the effective management to sensor resources is realized under measured data and reliable target prediction information. The results of simulations indicate that the method is effective and reasonable.

关 键 词：传感器管理 目标判定 支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]