1990 Fiscal Year Final Research Report Summary
Design of Robust Nonlinear Control System and Applications to Robot Multi-Link Mechanism Control
| Project/Area Number |
01550195
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
機械力学・制御工学
|
|---|
| Research Institution | Yamagata University |
|---|
Principal Investigator |
OKUBO Shigenori Institution, Department, Faculty of Engineering, Yamagata University, 工学部, 教授 (60134467)
|
|---|
| Project Period (FY) |
1989 – 1990
|
| Keywords | Nonlinear control system / Robust control / Robot manipulator / Parameter error / Coriolis force / Model following control system / Nonlinear compensator / Nominal value |
|---|
| Research Abstract |
At the present linear control theory has reached the complete level, on the other hand the developing speed of nonlinear control field is very low. All of actual physical system are nonlinear. For example tokamak type nuclear fusion system, mechtoronics system of robot manipulator, chemical plant of margarine production and so on are nonlinear systems. The conventional control methods for these systems are PID control or linear regulator system by using linearization around driving points. However the control law must be changed when the driving points or initial conditions are changed, so the function as control system are not enough. At this research project we developed a design method of model following control systems for general nonlinear systems and we could obtain robust stability conditions of which output errors converge zero and all of internal states become bounded. We applied above mentioned methods to the design for robot manipulators, and could get effective results. Sta
… More
te equations of robot systems can be obtained generally by using Lagrangian functions. Robot manipulators systems contain centrifugal force and Coriolis force which are complex nonlinear terms. These nonlinear systems can be described as an inner product of deterministic known functions and parameter vectors which contains observation errors. Control inputs can be constructed based on nominal values which can be got from observed parameters. This control scheme is a nonlinear compensator. The obtained systems are multi-dimention second degree systems which contain weak nonlinear terms, and have no zero point. This structure is proper on robot systems. For the characteristic performance of control system can be developed remarkably by model following control method.
Arbitrary pole assignment can be possible by the construction of linear model following control system to multi-dimention second degree systems to ignore the effect of weak nonlinear terms. Robust stability conditions can be obtained by solving a integral inequality for parameter errors with using Gronwall method. Also output errors can be reduced arbitrary small by sufficiently stable pole assignment of linear systems. We can get well results even if parameter errors are 50 percent by numerical simulations. Less
|
