ABSTRACT
The inverted pendulum (pole-cart balancing) problem has received a great deal of attention as a model problem for the establishment of learning control systems.1-7 That researchers are successful in this field can be seen in their published results. However, using a rigid pole as the pendulum, analysis shows that this system has only two degrees of freedom and little nonlinearity. As a result of these limitations, the learning controllers developed using such a demonstrator problem have limited power and are unlikely to have broad applications to manufacturing industries. Because of the limitations mentioned, this author modifies the pole-cart balancing problem to give a more exact test bed for learning controllers by replacing the rigid pole with an elastic pole. The dynamics of this new system are more complex and highly nonlinear when compared to the traditional rigid pole-cart balancing system as a result of the additional degree of freedom within the system, e.g., the transverse displacement of the elastic pole.20,24,28,37,42,43
Modeling and control of flexible robot systems has attracted much interest in recent years.8-14 This has arisen, in particular, in the area of space and industrial robots that require lightweight and flexible links. 15 Flexible robot manipulators have many advantages compared to robot manipulators constructed from rigid links. If the advantages associated with the lightweight machine elements are not to be sacrificed, then advanced control systems for such flexible robot manipulators have to be developed.9 The flexible pole test bed explored in this research allows the examination of some of the control issues within flexible linked robots.
