유전자 알고리듬 기반의 Input Estimation 방법의 개발 및 실험

Abstract

In this study, we develop the input estimation technique using the genetic algorithm to solve the inverse problem. We made the testbed consisting of beam structure with a laser sensor and accelerometer to validate our proposed input estimation technique. To solve the inverse problem, we need the exact parameter of the beam structure. To this end, the frequency response function was measured and the system identification was carried out. In the experiment, the beam structure was excited with specific disturbance. At the same time, the top displacement and acceleration of the beam structure were measured. The result of input estimation using the displacement is similar to the base input. However, if we use the acceleration data to estimate the input, the result cannot predict the low frequency vibration. In vibration system, the cause of the vibration is not low frequency but mainly high frequency disturbance. Therefore, we compared the estimated input from acceleration with base input filtered by low pass filter and both results were similar. From this result, it is confirmed that the input estimation using the acceleration can be applied to obtain disturbance information.