Research On Optimal Iterative Learning Control Algorithm And Its Applications

Iterative learning control is a technique for improving the transient response and tracking performance of processes, machines, equipment, or systems that execute the same trajectory, motion, or operation repetitively. The idea behind Iterative learning control is that if the system's operating conditions are the same each times it executes then any errors in the tracking response will be repeated during each operation. These errors and the control input signal can be recorded during each repetition and be uesed to modify the input signal that will be applied in the next repetition so that the tracking accuracy would be increased as the number of repetitions increases. With the emerging of the cyclic of repetitive systems or processes in industry, Iterative learning control has become a key technique of advanced controls.The convergence and convergence speed of conventional Iterative learning control algorithms are deeply affected by their learning coefficients. Take example by PID type iterative learning control algorithm, the selection of its learning coefficients can't benefit from the convergence analysis, but is completed according to experience. In order to avoid blindness in the selection of learning coefficients, some information about plant model is used to design learning control law. Thereby, a new method, the idea behind which is designing learning law by optimization, is educed and is called optimal iterative learning control algorithm. Optimal iterative learning control algorithm includes norm optimal iterative learning control algorithm and parameter optimal iterative learning control algorithm. Compare with norm optimal iterative learning control algorithm, parameter optimal iterative learning control algorithm based on norm performance is simpler, more efficient and can be realized more easily. However, there are some new questions need answer in the research on parameter optimal iterative learning control algorithm: 1) How to realize faster error convergence rate? 2) How to ensure the tracking error again converges monotonically to zero when the original plant is not positive? 3) A new kind of parameter optimal iterative learning control algorithm which can be used to deal with non- linear plant show be considered. 4) How to avoid the deep depend on plant model in optimal iterative learning control algorithms?In order to enhance learning efficiency and obtain faster and more accuracy transient tracking performances in iterative domain, a PID type fast parameter optimal iterative learning control algorithm based on norm performance index is founded in this paper. In the algorithm, the PID type operator is introduced to expend the dimension of the algorithm and to increase the free-degree of the optimal parameter.In this paper, a suitable set of basis functions is added into parameter optimal iterative learning control algorithm when the plant is not positive, and a new parameter optimal iterative learning control algorithm with basis function is proposed. Theoretic proof shows that the algorithm monotone convergence to zero no matter the plant is positive or not. In addition, simulations show that the algorithm also has a faster convergence speed compare with other similar algorithms.Clonal selection algorithm is improved and proposed as a method to solve optimization problems in iterative learning control, and a clonal selection algorithm based optimal iterative learning control algorithm is proposed in this paper. In the algorithm, more priori information was coded in the clonal selection algorithm to deal with constraints on input, at the same time, the size of the search space is decreased and the convergence speed of the algorithm is increased. In addition, real code and use of immune radius can make the optimal input obtained by the algorithm smooth.The superiority character of optimal iterative learning control algorithm is based on a precision model of plant, which usually can't be obtained in real condition. In order to cope with the uncertainty in the plant model, a clonal selection algorithm based model modifying device is introduced in the paper. After each trail, the input data and the error data between model output and real output are usaed to revise plant model, and the new plant model will be used in next trail. The model modifying device is designed for non-linear plant and also can be used in linear plant.Finally, 2-degree freedom"mass spring damping"active vibration control system is analyzed, and iterative learning control algorithms are used into the active vibration control. Compared to the traditional control methods, parameter optimal iterative learning control algorithm is a better choice especially at the low frequency. And during the variety of parameter optimal iterative learning control algorithms, the proposed PID type fast parameter optimal iterative learning control algorithm and the parameter optimal iterative learning control algorithm with basis function have better convergence character. Especially, the parameter optimal iterative learning control algorithm based on basis function, which can make the tracking error converge monotonically to zero, has a good futurity in active vibration control.