基于最小二乘支持向量机的测控数据融合  

Research on Fusion of Measurement and Control Data Based on Least Square-Support Vector Machine

作　　者：苏思[1] 姜礼平[1] 邹明[1] 

机构地区：[1]海军工程大学理学院,武汉430033

出　　处：《火力与指挥控制》2011年第3期98-100,114,共4页Fire Control & Command Control

基　　金：海军工程大学科研基金资助课题(hjsk200805)

摘　　要：提出利用最小二乘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,GPS数据经过时间配准处理与雷达数据达到时间同步,经过空间配准和坐标系变换,进行卡尔曼滤波,以滤波估计坐标值作为支持向量机的输入,以最小二乘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 least square-support vector machine 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 LS-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：最小二乘支持向量机 测控 数据融合 

分 类 号：TN957.52[电子电信—信号与信息处理]