Font Size: a A A

Research On Ship Course ADRC Combined With Local Approximation Neural Networks

Posted on:2018-10-16Degree:MasterType:Thesis
Country:ChinaCandidate:Y Z LiFull Text:PDF
GTID:2322330512477082Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
This paper mainly focuses on the design and simulation of large ship course controller which combined ADRC with local approximation neural networks.Both the mathematical model of the maneuvering of the large container ship and the algorithm design and Simulation of the intelligent controller are completed.Specifically as follows:Firstly,93888 tons displacement(9572TUE)container ship XIN SHANG HAI and 69500 tons displacement(5446 TUE)container ship SHANG HAI are selected as controlled plant in this paper,considering the interference of wind and current,the MMG model of ship maneuvering motion is established,then the model's accuracy and rationality are verified by completing turning test and zig-zag test.Secondly,a key feature of ADRC is that it can estimate the unknown disturbance as a state of expansion,and then implement the compensation to the system.So ADRC is applied to the course control system and the simulation is carried out on the basis of the theoretical design of ADRC and MMG model of XIN SHANG HAI.Thirdly,in order to improve the control effect of ADRC and the defects of too many parameters of nonlinearity ADRC,on the basis of studying two kinds of local approximation neural network theory of Cerebellar Model Articulation Controller(CMAC)neural network and Radial Basis Function(RBF)neural network,ADRC is improved by design two kinds of compound controllers,the first one is CMAC and ADRC compound controller CMAC-ADRC,the second one is RBF and ADRC compound controller RBF-ADRC.The simulation results of XIN SHANG HAI show that these two kinds of ship course ADRC combined with local approximation neural networks have greatly enhanced the control effect.Fourthly,SHANG HAI container ship is selected as another controlled plant in this paper and the control system simulations in different sea conditions are carried out.Then the applicability and effectiveness of the ADRC combined with local approximation neural network for the different scales and displacement ships are verified.
Keywords/Search Tags:Mathematical model of ship motion, ship course control, active disturbance rejection controller, Cerebellar Model Articulation Controller, Radial Basis Function neural network
PDF Full Text Request
Related items