Decoupling Control Of Multivariable System Based On Genetic Algorithm

With the rapid development of modern control technology,the control systems often have characteristics of multi variable,strong coupling and so on,which makes them more complex and difficult to be controlled with conventional control methods,and harder to achieve precise control,which presents a challenge in a wide range of industrial fields,such as electromechanical,communication,aerospace,etc.Focus on the coupling problem of multivariable system,several major decoupling methods were introduced,and the decoupling conditions was analyzed.In order to solve the problem in decoupling control of multivariable system such as the systems always have complex multi-parameter,high dimension,which leads to the difficulty in decoupling calculation to design the decoupling controller,a novel state feedback decoupling method based on genetic algorithm was proposed.The key lies in the design of appropriate parameters and operation mode of genetic algorithm,to ensure the algorithm works with preferably efficiency and gets suitable fitness function to evaluate the system’s coupling.Through the global optimization of genetic algorithm,the state feedback matrix and input transformation matrix were configured for the system,which are expected to make it approximate decoupling state continuously along with iterative calculation.Meanwhile,in order to ensure response performance of the decoupling system,the system’s output is evaluated with absolute error integral of step response,and then a criterion of integral square error-integrated time and absolute error is proposed to construct the fitness function of genetic algorithm,which can improve decoupling speed at the initial stage of the system response and reduce steady-state error at the later stage.Then,the proposed method was compared with classical decoupling algorithm by experiment,the results show that with genetic algorithm,the system can realize decoupling accurately,and moreover gets faster response speed,shorter regulation time.When it comes to practical application,the modeling and coupling analysis of a variable air volume air conditioning system were carried out,with genetic algorithm applied to its’ coupling control,expected results was obtained in experiments,verifying rationality and advantage of the proposed decoupling method.Finally,we discussed the decoupling problem of six axis force / torque sensor.In order to eliminate the coupling between axis,we collected the input and output sample data and system coupling information of the six axis force / torque sensor by experiments,then a decoupling method based on BP neural network optimized by genetic algorithm was proposed.By experiments,the optimization effect of genetic algorithm on BP neural network and the effectiveness of the proposed method were verified.