International Journal of Automation and Control http://www.inderscience.com/browse/index.php?journalID=109&year=2011&vol=5&issue=4 Inderscience Publishers Ltd en-uk International Journal of Automation and Control 1740-7516 1740-7524 © 2011 Inderscience Publishers Ltd editor@inderscience.com https://www.inderscience.com/images/files/coverImgs/ijaac_scoverijaac.jpg http://www.inderscience.com/browse/index.php?journalID=109&year=2011&vol=5&issue=4 http://www.inderscience.com/link.php?id=43610 Model predictive control (MPC) is one of the widely used concepts in the control engineering practice and in real-time applications. Generalised predictive control (GPC) is one of the most popular among the types of MPC. Because of its desirable features, it has become a popular tool for the control of most of the processes. However, there are certain limitations which restrict the use of GPC. In this paper, in addition to the basic concept of GPC, control law formulation, advantages and disadvantages, the other techniques that make use of constraints, continuous-time approach, providing delay, selecting sampled data and rate, use of delta operator, intelligent controllers, min max techniques with GPC which overcome the limitations of traditional GPC have been discussed. The review on GPC technique and its advances has been presented in this paper. A review on techniques with generalised predictive control
Kailas Holkar; Laxman Waghmare
International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 299 - 336
Model predictive control (MPC) is one of the widely used concepts in the control engineering practice and in real-time applications. Generalised predictive control (GPC) is one of the most popular among the types of MPC. Because of its desirable features, it has become a popular tool for the control of most of the processes. However, there are certain limitations which restrict the use of GPC. In this paper, in addition to the basic concept of GPC, control law formulation, advantages and disadvantages, the other techniques that make use of constraints, continuous-time approach, providing delay, selecting sampled data and rate, use of delta operator, intelligent controllers, min max techniques with GPC which overcome the limitations of traditional GPC have been discussed. The review on GPC technique and its advances has been presented in this paper.

]]> 10.1504/IJAAC.2011.043610 International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 299 - 336 Kailas Holkar; Laxman Waghmare Department of Electronics and Telecommunication, K.K. Wagh Institute of Engineering Education and Research, Panchavati, Nashik 422003, Maharashtra, India ' Department of Instrumentation Engineering, S.G.G.S Institute of Engineering and Technology, Vishnupuri, Nanded 431606, Maharashtra, India generalised predictive control GPC control law constraints model predictive control MPC. 2011-11-11T23:20:50-05:00 5 4 299 336 2011-11-11T23:20:50-05:00 http://www.inderscience.com/link.php?id=43611 This paper introduces a rotary double inverted pendulum (RDIP) systems. The model is derived by using Euler Lagrange. Linear quadratic regulator (LQR) controller is applied as the main controller to stabilise the rotary double. However, LQR alone cannot control RDIP efficiently because the plant derived in linear model is not exact model of the real plant. Practically, controller design aiming to guarantee robustness has to consider these uncertainties. In this paper, neural network predictive control is proposed to improve control performance of the conventional LQR controller. Results on control techniques from computer simulation are evaluated and compared. Hybrid NN predictive-based LQR controller for rotary double inverted pendulum systems: an analytical study
Viroch Sukontanakarn; Manukid Parnichkun
International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 337 - 355
This paper introduces a rotary double inverted pendulum (RDIP) systems. The model is derived by using Euler Lagrange. Linear quadratic regulator (LQR) controller is applied as the main controller to stabilise the rotary double. However, LQR alone cannot control RDIP efficiently because the plant derived in linear model is not exact model of the real plant. Practically, controller design aiming to guarantee robustness has to consider these uncertainties. In this paper, neural network predictive control is proposed to improve control performance of the conventional LQR controller. Results on control techniques from computer simulation are evaluated and compared.

]]> 10.1504/IJAAC.2011.043611 International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 337 - 355 Viroch Sukontanakarn; Manukid Parnichkun Faculty of Mechatronics, School of Engineering and Technology, Asian Institute of Technology, Pathumthani 12120, Thailand ' Faculty of Mechatronics, School of Engineering and Technology, Asian Institute of Technology, Pathumthani 12120, Thailand LQR linear quadratic regulator RDIP rotary double inverted pendulum NNPC neural network predictive control neural networks controller design simulation. 2011-11-11T23:20:50-05:00 5 4 337 355 2011-11-11T23:20:50-05:00 http://www.inderscience.com/link.php?id=43617 The robustness of back propagation neural network-based inverse model is researched by simulation in the paper. Firstly, the dynamic model of induction motor is built by state space theory. Secondly, the corresponding inverse model is got by inverse system theory. However, the analytic inverse model is hardly used in the engineering field by excessively depending on the parameters. Finally, an artificial neural network (ANN)-based inverse model, which synthesises artificial intelligent method and analytic method, is suggested. To accelerate the convergence of ANN and enhance its generalisation ability, the non-linear parts are realised by the analytic expressions and the corresponding results act as the inputs of ANN so that the complex-non-linear mapping relation become a simple-linear mapping and the structure of ANN is simplified. A three-layered feed-forward ANN with 12-10-2 structure is introduced to approach the inverse mode of induction motor drives. Finally, the robustness of ANN-based inverse model is verified by simulation in the case of parameter variance and state disturbance. Research on inverse model based on ANN and analytic method for induction motor
Shuo Ding; Qinghui Wu
International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 356 - 370
The robustness of back propagation neural network-based inverse model is researched by simulation in the paper. Firstly, the dynamic model of induction motor is built by state space theory. Secondly, the corresponding inverse model is got by inverse system theory. However, the analytic inverse model is hardly used in the engineering field by excessively depending on the parameters. Finally, an artificial neural network (ANN)-based inverse model, which synthesises artificial intelligent method and analytic method, is suggested. To accelerate the convergence of ANN and enhance its generalisation ability, the non-linear parts are realised by the analytic expressions and the corresponding results act as the inputs of ANN so that the complex-non-linear mapping relation become a simple-linear mapping and the structure of ANN is simplified. A three-layered feed-forward ANN with 12-10-2 structure is introduced to approach the inverse mode of induction motor drives. Finally, the robustness of ANN-based inverse model is verified by simulation in the case of parameter variance and state disturbance.

]]> 10.1504/IJAAC.2011.043617 International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 356 - 370 Shuo Ding; Qinghui Wu College of Engineering, Bohai University, Jinzhou 121013, China ' College of Engineering, Bohai University, Jinzhou 121013, China ANNs artificial neural networks inverse models decoupling control induction motor drives simulation dynamic modelling. 2011-11-11T23:20:50-05:00 5 4 356 370 2011-11-11T23:20:50-05:00 http://www.inderscience.com/link.php?id=43623 Artificial neural networks (ANNs) offer the versatility of being able to model the dynamics of a biosystem without requiring a phenomenological model. In addition, model predictive control (MPC) is a member of advanced discrete-time process control algorithms. The recent developments in the biotechnology due to MPC utilising the capability of ANN make the practical application of non-linear process control strategies a reality. This paper reviews the recent enhancement and applications of MPC in various biochemical processes using feedforward artificial neural networks which is also known as neural predictive control. The capability of neural predictive control to handle the common problems associated with biochemical processes, namely optimisation of objective function, optimisation of dynamic behaviour of the system, control of ill-defined non-linear systems, improving the computational efficiency of the strategy, disturbance rejection ability and evaluating the control effectiveness are discussed. The review clearly indicates that enormous work has been carried out involving dynamic behaviour of the bioreactor system which is analysed and optimised revealing that feedforward neural network has evolved as a good bioreactor neuro-controller. Feedforward artificial neural network to improve model predictive control in biological processes
Senthil Kumar Arumugasamy; Zainal Ahmad
International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 371 - 391
Artificial neural networks (ANNs) offer the versatility of being able to model the dynamics of a biosystem without requiring a phenomenological model. In addition, model predictive control (MPC) is a member of advanced discrete-time process control algorithms. The recent developments in the biotechnology due to MPC utilising the capability of ANN make the practical application of non-linear process control strategies a reality. This paper reviews the recent enhancement and applications of MPC in various biochemical processes using feedforward artificial neural networks which is also known as neural predictive control. The capability of neural predictive control to handle the common problems associated with biochemical processes, namely optimisation of objective function, optimisation of dynamic behaviour of the system, control of ill-defined non-linear systems, improving the computational efficiency of the strategy, disturbance rejection ability and evaluating the control effectiveness are discussed. The review clearly indicates that enormous work has been carried out involving dynamic behaviour of the bioreactor system which is analysed and optimised revealing that feedforward neural network has evolved as a good bioreactor neuro-controller.

]]> 10.1504/IJAAC.2011.043623 International Journal of Automation and Control, Vol. 5, No. 4 (2011) pp. 371 - 391 Senthil Kumar Arumugasamy; Zainal Ahmad School of Chemical Engineering, Universiti Sains Malaysia, Engineering Campus, Seri Ampangan, 14300 Nibong Tebal, Seberang Perai Selatan, Pulau Pinang, Malaysia ' School of Chemical Engineering, Universiti Sains Malaysia, Engineering Campus, Seri Ampangan, 14300 Nibong Tebal, Seberang Perai Selatan, Pulau Pinang, Malaysia artifical neural networks ANNs feedforward neural networks process control model-based control advanced control MPC model predictive control biotechnology optimisation modelling biochemical processes bioreactors. 2011-11-11T23:20:50-05:00 5 4 371 391 2011-11-11T23:20:50-05:00