Implementation of recurrent neural network to control rotational inverted pendulum using IMC scheme

Problem statement:This paper presents an overview of a controller for a Rotational Inverted Pendulum (RIP) based on a New Recurrent Neural Network (NRNN) using Internal Model control (IMC). The RIP consists of a DC servo motor, arm and pendulum. The RIP is modelled in MATLAB/Simulink and the simulation results are shown besides the experimental results. The proposed experiment shows intelligent method for stabilizing the RIP, which can recommend the control designers of nonlinear systems. The outcome exposed that the NRNN controller competent of controlling the RIP system productively, as exposed in the simulation results.

[1]  Ryohei Nakano,et al.  Stable behavior in a recurrent neural network for a finite state machine , 2000, Neural Networks.

[2]  Ben S. Cazzolato,et al.  Technical Report: On the dynamics of the Furuta pendulum , 2010 .

[3]  Kevin Warwick,et al.  Inverse model control using recurrent networks , 2000 .

[4]  Renato A. Krohling,et al.  Design of optimal disturbance rejection PID controllers using genetic algorithms , 2001, IEEE Trans. Evol. Comput..

[5]  Chang Wang,et al.  Design of blender IMC control system based on simple recurrent networks , 2009, 2009 International Conference on Machine Learning and Cybernetics.

[6]  M.O. Tade,et al.  Nonlinear model predictive control of a multistage evaporator system using recurrent neural networks , 2009, 2009 4th IEEE Conference on Industrial Electronics and Applications.

[7]  Rahib Hidayat Abiyev.  Recurrent Neural Network Based Fuzzy Inference System for Identification and Control of Dynamic Plants , 2007 .

[8]  R. Rahmani,et al.  A Hybrid Control Scheme for a Rotational Inverted Pendulum , 2011, 2011 UKSim 5th European Symposium on Computer Modeling and Simulation.

[9]  Chun-Jung Chen,et al.  Design of a Power System Stabilizer Using a new Recurrent Neural Network , 2006, First International Conference on Innovative Computing, Information and Control - Volume I (ICICIC'06).