摘要
为提高自主水下机器人(Autonomous underwater vehicle,AUV)的环境自适应能力,基于侧线感知机理,对水下目标进行形态识别和位置定位方法的研究。采用数值模拟的方法对流场中等边三角形的流场结构进行研究,提取"侧线"上的压力信号作为形态识别信息,训练并建立支持向量机(Support vector machine,SVM)识别模型。采用二步网格寻优的方法对SVM模型中的惩罚因子C和核函数参数g进行优化,模型测试表明,基于压力系数的时域波形结构可以对干扰源的形态进行辨识。通过提取压力系数波形中的特征值,对相对检测距离进行分析和拟合,结果表明利用压力信号的振幅,可以有效地计算出干扰源的相对位置。因此证明利用压力信号和SVM的方法,可以对水下目标进行识别和定位,为提高AUV的环境自适应能力提供了一种新思路。
In order to improve the environment adaptive ability of autonomous underwater vehicle(AUV),a method of form recognition and position location of underwater target is studied based on the lateral line sensing mechanism.The flow field structure of equilateral triangle is studied by numerical simulation.The pressure signal on the"lateral line"is extracted as identification information.And a support vector machine(SVM)recognition model is trained and established based on the data.The penalty factor and a kernel function parameter in SVM model is determined by two-step network method.The model test shows that the form of targets can be identified based on the pressure coefficient.The relative detection distance is analyzed and fitted by extracting the characteristic values of pressure coefficient waveform.The results show that the relative position of target can be calculated effectively based on the pressure amplitude.Therefore,it is proved that the pressure signal and SVM can be used to identify and locate underwater targets.The method provides a new idea for improving AUV environment adaptive ability.
作者
林兴华
武建国
王晓鸣
张敏革
刘海涛
LIN Xinghua;WU Jianguo;WANG Xiaoming;ZHANG Minge;LIU Haitao(School of Mechanical Engineering,Hebei University of Technology,Tianjin 300130;College of Mechanical Engineering,Tianjin University of Science and Technology,Tianjin 300222)
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2020年第12期240-248,共9页
Journal of Mechanical Engineering
基金
河北省自然科学基金(E2018202259)
河北省研究生创新(CXZZBS2017025)资助项目。
关键词
侧线系统
支持向量机
流场感知
数值模拟
目标识别
lateral line system
support vector machine
flow sensing
numerical simulation
target recognition
