Research Of Human-simulated Intelligent Control On Multi-axes Synchronous System

In manufacturing industry field, multi-axes synchronous system mainly includescnc system and manipulator. As a strong coupling, nonlinear, time-varying system,conventional PID control can’t meet the requirements of high speed and high precision.Therefore, a lot of improved control algorithms have been proposed in the last decades.Neural network control has strong fault-tolerance and the ability to approximatenonlinear functions, but exist self-learning problem. Fuzzy control don’t need precisemathematical model in the process of controller design, but it is hard to establish fuzzyrules. Adaptive control can change the parameters and structure of the controller toobtain good control results, but it can’t consideration to dynamic performance andsteady state performance. Sliding mode control has strong robustness, but existchattering phenomenon. Genetic algorithm can quickly deal with non-smooth even thediscrete problem, but its searching ability in new problems is limited. Iterative learningcontrol can achieve perfect tracking, but it is sensitive to initial value. In the controlfield of multi-axes synchronous system, these algorithms are better than PID control.However, all these algorithms exist the above defects, and some of these algorithmsdepend on system mathematical model.So far, human control performance is the most efficient control in the word.Human-simulated intelligent control adopts multi-mode control based on feature modeland feature identification, select different control modes according to differentcharacteristic state, so it can effectively simulate human control performance and obtaingood control performance. At the same time, the feature model established by systemerror and error derivative is insensitive to the change of system model, so the process ofcontroller design doesn’t depend on system model. In this paper, the concrete researchobject is manipulator. Firstly, analyze the manipulator’s kinematics and dynamics. Then,get the feature model by division the phase plane, and automatic correction controllerparameter in every concrete characteristic state, thus realize human-simulated intelligentcontrol in manipulator. Lastly, verify the effectiveness of human-simulated intelligentcontroller by compared with PD controller on simulation platform...