Chaotic signal design for frequency domain identification

System identification usually is the first step to understand control system and to design controllers, especially where performance matters, such as high speed and high precision motion control system. Theoretically, there are certain conditions of the excitation signals. For example, the white noise is one such type of signal. However, some modes are not allowed to be excited in practice. On the other side, the input signal to the mechatronic system contains no high frequency harmonic components at the working mode. So it is of its differential. Based on these observations, a particular class of chaotic signal is proposed as the excitation signal for identification experiments. The model of the system is obtained by least squares method. The simulation and experimental results show that such chaotic signal can be applied to mechatronic system safely. Further, the identified model can approximate the model of the system, especially on the frequency range of concerned.