This thesis studies a semi-active suspension system for improving a ride comfort. A quarter bus is modeled with MSC.ADAMS. The quarter bus computational model is composed of a sprung mass, air spring, passive damper, variable damper, unsprung mass, and tire. A test data is acquired from a part test of air spring and passive damper. The test data showing nonlinear characteristic is represented by a spline function. Then, the spline function is included in the bus computational model.

A controller for control the semi-active suspension is developed with MATLAB/Simulink. The controller is composed of a main controller and assistant controller. The main controller is developed using a fuzzy control. The assistant controller is developed using a self-tuning fuzzy control. The main fuzzy controller uses a relative velocity between sprung mass and unsprung mass, and a absolute velocity of sprung mass as input variables. A damping force of a variable damper is used as output variable. A triangular shaped function is used as a membership function for input and output variables. Input variable is represented by five membership functions. Output variable is represented by seven membership functions. 25 control rules are designed for the main fuzzy controller. Input variables for the self-tuning fuzzy controller are selected as the main fuzzy controller. A weight is selected as output variable for the self-tuning fuzzy controller. The weight is multiplied to the output of the main fuzzy controller. A gaussian function is used as a memebership function for input and output variables of the self-tuning fuzzy controller. Input variable is represented by three membership functions. Output variable is also represented by three membership functions. 9 control rules are designed for the self-tuning fuzzy controller.

MSC.ADAMS/Control is used for interfacing the bus computational model and the controller. In order to show the performance of a semi-active suspension system using a self-tuning fuzzy controller proposed in this paper, a simulation is performed on a road with a random profile, and a bumpy road. A driving simulation on a road with a random profile is performed with a velocity of 70km/h. A driving simulation on a bumpy road is performed with a velocity of 30km/h. The performance of a semi-active suspension system using a self-tuning fuzzy controller proposed in this paper is compared with the one acquired from a passive suspension, active suspension, and sky-hook suspension.