Real-time intelligent controllers for systems with partially known dynamic models

In this paper, the strength of two different control theories were combined to construct an intelligent control system. The modern control theory is first used to design a full-state feedback controller. This controller provides the needed accuracy of the control system performance at a specific operating range (modeled range) of the plant dynamic. Then the fuzzy logic technology is used to design a bank of controllers, which handle the intelligent decision-making of the control system at the remaining operating range of the plant dynamics. Typically this is the un-modeled dynamics of the controlled plant. The developed intelligent control system was used to balance inverted pendulum hardware. The controller was first simulated then implemented and tested on the dSPACE real-time controller board. The results showed that the intelligent control system could provide a practical control performance for the inverted pendulum hardware starting from any arbitrary initial condition including failed conditions with unexpected external disturbance.