摘要
现代战争越来越多地使用高科技武器,各种武器车辆坦克的声音相互混叠,进行声目标识别是极有挑战性的。为了在嘈杂的战场环境中多传感器网络仍能正确识别各种目标信号,结合无线多传感器网络的特点和数据融合理论,设计出了适合于战场环境的声目标识别算法。使用小波包进行预处理及特征提取,人工神经网络进行分类识别,并运用数据融合算法得出最终识别结果。通过对采集到的声目标信号进行识别,结果表明方法应用于战场声目标识别中是可行有效的。
The high - tech weapons are used more and more in modern war, and various voices of vehicles and weapons make the environment noisy, so it is a challenge to identify acoustic objective in such a noisy situation. An acoustic target recognition algorithm is given based on the characteristics of wireless sensor networks and the data fusion theory so that the various target signals can be identified correctly by wireless multi - sensor network in a noisy battlefield environment. The algorithm uses wavelet packet pretreatment and feature extraction, artificial neural net-work classification, and the data fusion algorithm to work out the final results. The result shows that the algorithm is feasible and effectual for battlefileld acoustic target identification through identifying original acoustic signal.
出处
《计算机仿真》
CSCD
2008年第10期1-4,26,共5页
Computer Simulation
基金
国家自然科学基金(60472074)
关键词
声目标识别
多传感器网络
特征提取
数据融合
神经网络
Acoustic target identification
Multi - sensor network
Feature extraction
data fusion
BP network