基于支持向量机的数据融合机动目标跟踪算法  

A Data Fusion Algorithm for Maneuvering Target Tracking Based on SVM

作　　者：苏思[1] 姜礼平[1] 刘惊涛[2] 赵大川 

机构地区：[1]海军工程大学理学院,武汉430033 [2]河南大学郑州分校,郑州450052 [3]中原出版传媒集团,郑州450016

出　　处：《河南大学学报（自然科学版）》2010年第3期299-302,共4页Journal of Henan University:Natural Science

基　　金：海军工程大学科研基金资助课题(hjsk200805)

摘　　要：利用支持向量机方法研究GPS和雷达系统对机动目标联合测量中的数据融合问题.使GPS数据经时间配准处理与雷达数据达到时间同步,再经空间配准和坐标系变换后进行卡尔曼滤波,将滤波估计坐标值作为支持向量机的输入,以支持向量机为同步融合中心,输出为目标轨迹的融合估计值.仿真结果表明,这种方案可以达到比融合前数据更贴近真实值的效果.A SVM data fusion approach for GPS and radar system＇s joint observation of maneuvering target tracking was presented. After time registration,the measurements from GPS would keep synchronous with the radar measurements,then the steps of sensor registration,coordinate conversion and Kalman filtering were taken. The processed data were then transmitted to the synchronous SVM fusion center as the input data,the output data were considered as the estimated coordinates of the target. Simulation results showed that this algorithm is effective to improve the processed data＇s precision and stability on the whole,with less amount of training samples than neural network algorithm.

关 键 词：支持向量机 数据融合 目标跟踪 

分 类 号：TN957.52[电子电信—信号与信息处理]