Research Of Multi-Pendulum System Human Simulated Intelligence Motion Control Based On Sensory-Motor Schema

Multi-pendulum system is high-rank, multi-variable, nonlinear, strong-coupled and underactuated inherent unsteady one, and its control process includes almost all typical control problems. As a kind of experiments equipment, pendulum system is also very cheap and is a simple structure, but its control effect is clear at a glance. Therefore, swinging up and handstand control of multi-pendulum is used to validate the effectiveness of different control theories and methods. Spoken from motion control, multi-pendulum is also quite similar to multi-finger manipulator, wheeled mobile robot, humanoid robot, space robot etc. In the systems, the coupling between joints is always utilized to complete different control target, and the constraints brought by the coupling must be avoided as possible. Because there exist underactuated joints, multi-pendulum system belongs to second unholonomic ones. As a result, most common continuous feedback control methods become invalid to control these systems. Therefore, research on motion control of multi-pendulum system has not only important meaning for the study of human-simulated or biology-simulated mechanisms, but also significant value as reference to improve the control performance of robot systems.The motion control of multi-pendulum system is a typical problem in the control field, and these corresponding real-time controlling results are presented very little in both here and abroad. Moreover there lack an integrated control theory or method to resolve these problems. Different from commom motion control, humankind's or other biologies'sensor-motor intelligences are based on the structure of multi-mode and multi-controllers. They are good at utilizing close cooporation of these controllers and controlling modes, and consequently complete all kinds of complex control tasts by the time-space coordinations and conflict mediation. Besides, human does also well in the study and optimization according to the characteristics of tasts, environments and objects. Compared with existing controllers,"human controller"possesses higher static and dynamic performance.Under the direction thought of"human-simulated"and"biologies-simulated", the paper combined many-year fruit of human-simulated intelligent control with schema theory in cognitive science and productions of genetic algorithm, simplified human sensor-motor intelligences to 3C(complex plants, complex mission, complex circumstance) as the decomposition of complex tasts, the building of objects'model...