| Summary: | 碩士 === 淡江大學 === 資訊工程研究所 === 81 === In modern industrial control application there are many
situations in which the control objective is to move actuator
from one point in space to another and to settle as quickly and
accurately as possible at the point, while the control effort
is kept within some limited value. Therefore , it is usually
necessary to design two tracking controllers for the sysetm to
accomplish large movements. During the first stage, when the
actuator is undergoing large accessing motion, the control
algorithm may be designed by using the classical time optimal
control with constraint the control effort. As the actuator
comes close to target position , the controller is switched to
a regulation stage, The method is to eliminate the chattering
and steady state error common to the classical time optimal
Bang-Bang controller. In this paper a new method for nonlinear
saturated control is proposed. The method is based on the
classical time optimal Bang-Bang control. By preserving
the time optimal characteristics of the classical Bang-Bang
control, the chattering and steady offset problem will be
avoided by the new method. A fuzzy controller is utilized for
the controller implememtation. The method is inherently
robust. The proposed controller "Optimal Path Estimation Method
(OPEM)" emerges from the concept of sucessively suppressing
dimensions of the state space and decreasing control effort
value. The feedback signal for the fuzzy logic controller is
based on the distance from the present state to the switching
curve along the optimal trajectory, and from the switching
curve to the original point of phase plan. The advantages of
proposed method over traditional Bang-Bang control method are
illustrated by numerical simulations of a servomotor system.
|
|---|
