Research On Initial State And Delaying In Iterative Learning Control

Iterative learning control is a simple algorithm and it can constantly improve thetracking performance of the system output by constantly correcting system controlinput. Because of its constantly revised performance, the system is able to achievehigh-precision tracking to any given desired trajectories in the limited time.Meanwhile iterative learning control has some advantages such as it does not dependon accurate mathematical model of the system, and it does not need the system’s priorknowledge, and it does not need identifying the parameters of the system model, and itis simple when computation. These advantages improve the ability of controllingmotion system with characteristics of nonlinear, difficult modeling, strong coupling,complex and so on. Iterative learning control is a high-precision control method, andits tracking accuracy can continuously improve with the increasing of iterations.Iterative learning control is widely researched because of these advantages.Firstly, this paper mainly analyses the arbitrary initial states and delaying on thebasis of the current studies, and concludes that arbitrary initial states and delayingfactors have an important impact on the performance of the system.Secondly, this paper studies the initial state of iterative learning control is anyvalue and then this paper puts forward a new iterative learning algorithm, which canconverge to the desired initial state. If the initial state is not equal to the desired initialstate in the process of iteration, the system cannot achieve no difference tracking, andit can only achieve poor tracking, at the same time the system is unstable. While theiterative learning algorithm the paper proposed can adjust the system state to the initialstate of the system expects in any case. The system can achieve error-free tracking ifgo on iterative learning controlling. The form of the algorithm is simple, and it is easyto implement. Finally it shows that this algorithm’s control effect is good by itssimulation results.Finally, this paper proposes a new iterative learning algorithm for a class ofnonlinear systems, which can achieve high-precision tracking. This paper designs a new iterative learning algorithm according to the control delaying and state delayingfor the nonlinear systems. The algorithm does not require precise state delaying and itcan ensure that the system output strictly track the desired trajectory as long as we candetermine its boundary value. Finally it verifies the feasibility and effectiveness of thealgorithm by theoretical proving and simulation studies.