Research On Time Delay Compensation Of Teleoperation Robot Based On Neural Network Prediction

With the rapid development of network communication technology,remote control of robots via the Internet has become an important research direction,and the application of network teleoperation technology is becoming more and more extensive.The teleoperation robot greatly extends the control distance,which can replace human beings to complete operations in extreme distances and harsh environments.It has demonstrated its superiority in telemedicine,space exploration,and dangerous rescue.Because in the remote control,there are problems such as random delay and packet loss in the network,which will reduce the running accuracy of the robot and affect the system stability,it is of great significance to eliminate the influence of time delay on the network control system of teleoperation robot.Therefore,in view of the delay problem in the network control system of teleoperation robot,this thesis combined with the improved WPA-BP neural network delay prediction,designs a delay controller based on the generalized predictive control algorithm,and builds a remote control platform for practical verification.The main content of this thesis includes the following aspects:Firstly,the structure of teleoperation robot system based on C/S mode is designed.The mathematical model of the manipulator is established.It is implemented according to the mathematical model in Sim Mechaincs toolbox.The simulation platform of network control system is built by using True Time toolbox combined with manipulator model,and the problems of the network control system are analyzed.For the actual network delay,the Ping program was used to obtain the delay.Secondly,a BP neural network delay prediction model based on improved wolf group algorithm optimization is designed.Because the design of the controller requires the use of network delay,the BP neural network model is used to predict the actual delay.For the BP algorithm,the convergence speed is slow and it is easy to fall into the local extremum.The wolf group algorithm is used to optimize the BP weight and threshold.and the behavior of the wolf group algorithm is improved.The improved WPA-BP delay prediction model is established,which realizes the effective prediction of the actual delay.Then,a controller based on GPC algorithm is designed.The GPC algorithm is designed for the linearized manipulator dynamics model,and the GPC controller is designed based on the improved WPA-BP prediction delay.The adaptive prediction step size and corresponding control strategy are proposed,and the network control simulation platform is built.The delay controller has better control effect.Finally,a remote control platform for teleoperation robot is built.Using TCP/IP protocol and socket programming,the communication system is designed.The Microsoft Visual Studio 2010 tool is used to design the client and server control platform.VC++ is used to program and control algorithm is written into the software,which realizes the remote control of the mechanical arm.The actual test of the whole system proves that the control algorithm can better compensate the random delay of the network and reduce the influence of delay on the mechanical arm system.It has important applications in the field of robot remote control such as space exploration and dangerous search and rescue.