【摘要】 若一般的控制系统产生非线性故障时 ,此系统可能会出现混沌运动。对高维系统的测量得到的通常只是一维时间序列。该文利用RBF神经元网络和相空间重构理论 ,由测量所得的实验数据序列计算系统的Lyapunov指数谱。RBF网络的中心的选取采用自组织学习算法。对Henon映射的仿真结果说明 ,此种方法的正确性以及计算所需样本点较少 ,计算速度快 ,精度较高。对飞行仿真转台的具体实验数据的仿真结果表明 ,此方法具有重要的实际应用意义更多还原

【Abstract】 The control system is probably chaotic, when it has the nonlinear obstacles. When we measure the high dimension system, we usually acquire the one-dimension time series of this system. We estimate the Lyapunov exponents for this system through the theory of RBF neural network and phase space reconstruction, using the time series of this system. We use the self-organize algorithm to find the location of RBF centers. The computing results of the Henon map show that this method has higher precision and computing speed. The computing results of the specific data from the flight simulator show that this method has practical significance.更多还原