Research On Iterative Learning Control Method Using Trajectory Primitive Matching

With the application of industrial robots in automatic manufacturing system,people have a high demands for effective control when a task is changing rapidly.For the precision machining tasks,iterative learning control(ILC)is put forward.Through repeated work,the tracking error will decrease along the entire operating range using ILC.However,it depends on the repeatability of reference trajectory,interference and dynamic system process that ILC achieve tracking reference trajectory completely.Once the task(the reference trajectory)changes,it must to learn again because the accumulated learning control signals in the early stage will be null.Therefore,for the problem that ILC must relearn when the task changes.And an ILC algorithm is proposed based on the optimal trajectory primitive matching method in this thesis.The thesis's main results consists of the optimal trajectory primitive matching method,the control method of iterative learning feedforward and H_? feedback,extracting the initial iteration control signals and so on.The main work of this thesis is described as follows:(1)The background on the subject is introduced briefly,and the research status of ILC and H_? control are reviewed.(2)The overall architecture of the system is introduced that ILC based on the optimal trajectory primitive matching method,and the system block diagram is given.Then,the method of extracting initial iteration control signals is introduced.(3)The optimal trajectory primitive matching and combination algorithms are studied on.Including establishing a NURBS model of time and space,a similarity judgment method of two segments curves based on Kabsch algorithm,the optimal matching and combination algorithm based on the given trajectory index of similarity and minimum search step.And then the results of simulation experiment are given.(4)The ILC feedforward and H_? feedback control method is studied on.There is the deviation between the initial control signals and expective initial control signals.The deviation is regarded as the interference of the system.So the H_? control method is used to restrain the effects of interference on the system.Finally,through the simulation experiment,the results of the control method are analyzed.(5)A two-link mechanical arm model is established,which has compensated for its nonlinear part.Then the experimental results are given to verify the effectiveness of the above methods.Finally,the innovation and the shortage is summarized,and some contents needing to be improved are put forward.