NEW LEARING CONTROL SCHEMES FOR NON-UNIFORM TRAJECTORY TRACKING WITH NON-PERIODIC SYSTEM UNCERTAINTIES

Abstract

Learning Control is an effective, yet simple technique for dealing with repeatable control problems. The main concern of this thesis is to develop new learning control methods for dealing with non-repeatable process as well as task uncertainties and consequently widen the application range as well as improve control performance of learning control. Direct learning control schemes have been developed recently to solve non repeatable trajectory tracking problems. Unlike conventional iterative learning schemes, DLC schemes learn a set of unknown basis function vectors which can be used to generate the desired control profile of a new trajectory. A drawback of DLC is that the inverse matrix calculation is inevitable, which is time consuming and may result in singularities due to the batch processing nature. A Recursive Direct Leaming Control method is proposed which learns the basic function vectors and overcomes the implementation difficulties in DLC schemes. The focus of this work is on learning the control profile of trajectories with same operation period but different magnitude scales. Non-repeatable uncertainties in the plant dynamics are considered in the next part of the work. An inverse dynamics partition separates the repeatable and nonrepeatable components of the control input. A Robust Learning scheme is synthesized for learning the repeatable components and suppressing the non-repeatable components by variable structure control. Since robust control requires the bounding function knowledge of uncertainties, adaptive robust control is used to suppress the non-repeatable components. Based on the ability to learn repeatable components and reject non-repeatable components an integrated learning control scheme is developed to handle nonrepeatable trajectory tracking problems. The integrated learning control scheme learns the unknown basis function vectors (estimated in RDLC using previous control knowledge) over the iterations and suppress the non-repeatable uncertainties using an adaptive robust control approach. Simulation results are provided to confirm the features of the proposed algorithm.