中文摘要：

针对武器自动机含噪冲击振动信号的混沌特性识别问题,提出了基于神经网络的混沌特征识别算法。将信号进行降噪处理,利用神经网络强大的学习和非线性处理能力,逼近信号真实相空间映射建立Jacobian矩阵,通过Jacobian矩阵计算出最大Lyapunov指数,并判断信号是否含有混沌特征。分别采用Lorenz仿真系统和自动机动作的冲击振动实测信号进行算法验证,仿真和试验结果表明提出的算法可以有效地解决自动机冲击振动信号混沌特性识别问题。

英文摘要：

In response to the issue of identifying the chaotic characteristics of automatic gun impact vibration signals,this paper puts forward an identification caculation method based on the BP neural network. Firstly,a method is used to reduce the noise in the signal measured. Secondly,based on the robust nonlinear reflection and training function of artificial neural networks,the optimal direction estimation of the signal real map can be obtained to create the Jacobian matrix from the output of the neural network,for the method of the Jacobian based approach can estimate the maximum Lyapunov exponent and judge if the signal has chaos. Positive maximum Lyapunov exponent was obtained from the signals,showing certain chaos features. Both the chaotic Lorenz simulation signal and the automatic gun impact vibration signal are respectively used for identitying the chaos property with the proposed method. Simulations and experiments verified the validity of the proposed method.