The dynamic characteristics of a special weapon vehicle moving on a random road profile is investigated. An optimal controller is introduced to the vehicle suspension system in an effort to reduce the vibration level. And the resulting stochastic control system is simulated to see the effect of the control system on the vibration of the vehicle. First, the vehicle system is modeled as a two dimensional system with eight degrees of freedom. And this model is further simplified to two 2-dimensional systems in order to avoid the mathematical complication involved in the original model. Also two simplified 3-dimensional models are derived for 3-dimensional analysis of dynamic characteristics. The equation of motion for each model is derived and put into state variable form. By the state variable technique, power spectral responses of the bouncing, pitching and rolling motion are obtained. And mean squre responses are obtained through integrations of power spectral densities. Those results for each model are discussed and compared. The results show that the response characteristics of the simplified models are in excellent agreements with those of the simplified models for the vibration control system design of the vehicle. An optimal controller introduced to the suspension system is so chosen that the integral quadratic cost functional is minimum, and the state variable feedback law is thus obtained. The controller utilizing the information of the full state variables is assumed to be inaccessible to some state variable, and also sensing noise of the measurable variables is assumed to be present. Under these situations the Kalman filter algorithm is incorporated with the feedback law. The results obtained from the control system simulations indicate that with an approiate displacement and velocity control the vehicle vibration reduced effectively. Also the effectiveness of this control in the vibration reduction is found to be more sensitive to the variation of the number of the sensed state variables than to the variation of sensing noise.
특수차량이 Random한 도로 형상위를 주행할때 나타내는 동적특성을 알아보고 문제의 간단화를 위해 진동System의 수학적 모형의 간단화를 시도하였다. 또한 진동특성을 Control 할수있는 Control System을 설치하는데 현대 Control개념과 estimation technique 를 적용하는 문제를 다루었다.
바퀴의 질량을 포함한 상태에서 간단화된 B,D - 모형은 power spectral density상의 low frequency에서, 바퀴의 질량을 무시한 상태에서 간단화된 C,E - 모형은 mid-frequency에서 간단화 되기전 모형과 잘 일치했고 mean square값은 각 모형들이 모두 5% 이내의 오차를 보여 주었다.
각 바퀴가 겪는 도로 형상 사이의 Correlation을 고려해준 경우에 비해 고려해 주지 않은 경우가 Bouncing과 Rolling Motinn의 mean square값은 $\frac{1}{6}$ 정도 감소하고 Pitching은 2.5배 정도 증가한 값을 보여주고 있다.
후반부에서는 Control System Simulation결과에서 변위Control을 하기 위해서는 속도의 적당한 Control이 함께 필요한 것을 볼수 있었고 측정error보다는 측정상태변수의 수가 Control효과에 미치는 영향이 크게 나타났다. 따라서 이러한 multi-variable System에서 보다 정확한 estimation을 위해 보다 많은 상태변수를 측정하는 것이 필요하다는 것을 알수 있다.
