Industrial automation systems have a hybrid nature. Such systems contain continuous activities an... more Industrial automation systems have a hybrid nature. Such systems contain continuous activities and discrete event activities, interacting with each other. This paper proposes a control algorithm dealing with the continuous activities, as well as the discrete ones. The discrete event controller performs resource allocation and coordination tasks at the higher level of the hybrid system. The direct implementation of the
International Journal of Control, Automation and Systems, 2013
ABSTRACT Early fault detection and isolation minimize cost and processing time of industrial syst... more ABSTRACT Early fault detection and isolation minimize cost and processing time of industrial systems. The main purpose of this paper is to introduce a novel delta-marking idea to design a generalized interpreted Petri net-based fault detection and isolation scheme (IPN-based diagnoser) for all PN models. This idea overcomes the event detectability problem that restricts the performance of the conventional dignosers to certain types of Petri net model. Furthermore, a unified framework for a discrete event system that comprises a PN-based model, a PN-based supervisor, and a PN-based diagnoser of a process under consideration, is built in this paper. In addition, this paper addresses some issues for industrial systems modeling and supervision. A rapid thermal process (RTP) is used as an industrial process to test the proposed unified framework after revising its PN-model. This scheme is not only developed for RTPs, but also can be employed for industrial processes with certain modifications depending on the nature of their structures. Simulation results have confirmed the effectiveness of the proposed unified framework.
Developing techniques for mobile robot navigation constitutes one of the major trends in the curr... more Developing techniques for mobile robot navigation constitutes one of the major trends in the current research on mobile robotics. This paper develops a local model network (LMN) for mobile robot navigation. The LMN represents the mobile robot by a set of locally valid submodels that are Multi-Layer Perceptrons (MLPs). Training these submodels employs Back Propagation (BP) algorithm. The paper proposes the fuzzy C-means (FCM) in this scheme to divide the input space to sub regions, and then a submodel (MLP) is identified to represent a particular region. The submodels then are combined in a unified structure. In run time phase, Radial Basis Functions (RBFs) are employed as windows for the activated submodels. This proposed structure overcomes the problem of changing operating regions of mobile robots. Read data are used in all experiments. Results for mobile robot navigation using the proposed LMN reflect the soundness of the proposed scheme.
This paper introduces a novel neural network named Auto Regressive eXogenous Local Model (ARX-LM)... more This paper introduces a novel neural network named Auto Regressive eXogenous Local Model (ARX-LM) network. It basically adopt the philosophy of forming a process be modeled with a set of fuzzy-wavelet submodels. Each submodel comprises a Takagi-Sugeno-Kang (TSK) fuzzy model fertilized by a wavelet function. The outputs of these submodels are weighted and summed to produce the final output. The main feature of the proposed LM network is that its structure is an ARX model that inherits the simplicity of the conventional linear and non-linear control systems, the learning capabilities of neural networks, the transparency of fuzzy systems and the locality of wavelet networks. The former means that the Lyapunove’s direct method can be employed to check the stability of the proposed ARX-LM network. The latter three features make the proposed network, powerful, transparent, and plastic. The plastic feature signifies that a new pattern can be assigned to uncommitted cluster if this pattern ...
Abstract: The Genetic Algorithm (GA) has the superiority over classical optimization algorithms i... more Abstract: The Genetic Algorithm (GA) has the superiority over classical optimization algorithms in finding the optimal (global) solution in multi-parameter search space. Because it simultaneously evaluates many points in the parameters space, it is more likely to converge toward the global solution. It need not assume that the search space differentiable or continuous and can also iterate several times on each datum received. This chapter proposes two genetic-based Takagi-Sugeno-Kang Recurrent Fuzzy Neural Network (TSK-RFNN) schemes; one is TSK-RFNN-G and the other is TSK-RFNN-GMRC. The former employs the well-known GA while the latter introduces the Model-Reference Control (MRC) to the former. The proposed algorithm uses real data instead of binary data to represent a chromosome. Unlike the TSK-RFNN-G, the TSK-RFNN-GMRC selects the controller’s parameters online (at each time step). For comparison reasons, both the proposed schemes were tested for controlling time intensive process...
Abstract: Elman network is a class of recurrent neural networks used for function approximation. ... more Abstract: Elman network is a class of recurrent neural networks used for function approximation. It has a set of global sigmoid functions at its hidden units. That means that if the operating conditions of a process be identified, are changed the function approximation property of the network is degraded. This is due to the fact that the universes of discourse of the network is covered by global sigmoid functions. This paper introduces a new version of the Elman network named Elman Recurrent Wavelet Neural Network (ERWNN). It merges the multi-resolution property of the wavelets and the learning capabilities of the Elman neural network to inherit the advantages of the two paradigms and to avoid their drawbacks. Structurally, the number of hidden/context units of an ERWNN should at least be equal to the order of the system be modeled. Stability and convergence property is proven for the proposed network. The dynamic back propagation (DBP) algorithm is employed to train the proposed ne...
Abstrncf - The performance of a model-based control system depends strongly on the accuracy of th... more Abstrncf - The performance of a model-based control system depends strongly on the accuracy of the process model used. This paper proposes a hybrid wavelet fuzzy neural network for modeling and control purposes. The proposed network inherits the advantages of both fuzzy and ...
The Genetic Algorithm (GA) has the superiority over classical optimization algorithms in finding ... more The Genetic Algorithm (GA) has the superiority over classical optimization algorithms in finding the optimal (global) solution in multi-parameter search space. Because it simultaneously evaluates many points in the parameters space, it is more likely to converge toward the global solution. It need not assume that the search space differentiable or continuous and can also iterate several times on each datum received. This chapter proposes two genetic-based Takagi-Sugeno-Kang Recurrent Fuzzy Neural Network (TSK-RFNN) schemes; one is TSK-RFNN-G and the other is TSK-RFNN-GMRC. The former employs the well-known GA while the latter introduces the ModelReference Control (MRC) to the former. The proposed algorithm uses real data instead of binary data to represent a chromosome. Unlike the TSK-RFNN-G, the TSK-RFNN-GMRC selects the controller’s parameters online (at each time step). For comparison reasons, both the proposed schemes were tested for controlling time intensive processes e. g. th...
Elman network is a class of recurrent neural networks used for function approximation. The main p... more Elman network is a class of recurrent neural networks used for function approximation. The main problem of this class is that its structure has a set of global sigmoid functions at its hidden layer. That means that if the operating conditions of a process be identified, are changed the function approximation property of the network is degraded. This paper introduces a new version of the Elman network named Elman Recurrent Wavelet Neural Network (ERWNN). It merges the multi- resolution property of the wavelets and the learning capabilities of the Elman neural network to inherit the advantages of the two paradigms and to avoid their drawbacks. Stability and convergence property is proven for the proposed network. The paper also develops a model reference control scheme using the proposed ERWNN. The proposed scheme belongs to indirect adaptive control schemes. The dynamic back propagation (DBP) algorithm is employed to train both the two networks structured for the indirect control sch...
Recently, an efficient method for obtaining reduced interpreted Petri net, IPN, diagnoser was pre... more Recently, an efficient method for obtaining reduced interpreted Petri net, IPN, diagnoser was previously proposed for discrete event systems. This diagnoser consists of a single place and the same number of transitions that the system model has. Although, the current marking of this place (diagnoser) is enough to determine and locate faults occurring in a simple diagnosable discrete event system, it fails to determine and locate faults occurring in a complex non-diagnosable hydride systems that have discrete and continuous variables. Diagnosability means that the PN model is live, strongly-connected, and Transition-invariant (T-invariant); however, most of PN models are not diagnosable. This paper generalizes the IPN-based diagnoser to deals with such complex systems with ease. Chemical batch processes are employed to test the modified IPN-based diagnoser and the simulation results obtained indicate that the proposed diagnoser is promising for industrial processes.
Elman network is a class of recurrent neural networks used for function approximation. The main p... more Elman network is a class of recurrent neural networks used for function approximation. The main problem of this class is that its structure has a set of global sigmoid functions at its hidden layer. That means that if the operating conditions of a process to be identified, are changed the function approximation property of the network is degraded. By noting the close analogy between neural networks, we may conclude that cross-fertilization between two different technologies may result a more powerful new one. Thus, this paper introduces a new version of the ELMAN network named ELMAN Recurrent Wavelet Neural Network (ERWNN). It merges the multi- resolution property of the wavelets and the learning capabilities of the ELMAN network to inherit the advantages of the two paradigms and to avoid their drawbacks. Stability and convergence property is proven for the proposed network. The dynamic back propagation (DBP) algorithm is derived to train the ERWNN. This paper also develops a model ...
2006 IEEE International Conference on Fuzzy Systems
Auto Regressive Moving Average (ARMA) models have attractive properties of simple computations an... more Auto Regressive Moving Average (ARMA) models have attractive properties of simple computations and fast learning. However, the unknown structures of the processes to be modeled produces difficult harmony between them and their ARMA models. This paper develops a novel ARMA model using the Particle Swarm Algorithm (PSA) to overcome this discord problem. It also introduces a PSA-based fuzzy control scheme
Industrial automation systems have a hybrid nature. Such systems contain continuous activities an... more Industrial automation systems have a hybrid nature. Such systems contain continuous activities and discrete event activities, interacting with each other. This paper proposes a control algorithm dealing with the continuous activities, as well as the discrete ones. The discrete event controller performs resource allocation and coordination tasks at the higher level of the hybrid system. The direct implementation of the
International Journal of Control, Automation and Systems, 2013
ABSTRACT Early fault detection and isolation minimize cost and processing time of industrial syst... more ABSTRACT Early fault detection and isolation minimize cost and processing time of industrial systems. The main purpose of this paper is to introduce a novel delta-marking idea to design a generalized interpreted Petri net-based fault detection and isolation scheme (IPN-based diagnoser) for all PN models. This idea overcomes the event detectability problem that restricts the performance of the conventional dignosers to certain types of Petri net model. Furthermore, a unified framework for a discrete event system that comprises a PN-based model, a PN-based supervisor, and a PN-based diagnoser of a process under consideration, is built in this paper. In addition, this paper addresses some issues for industrial systems modeling and supervision. A rapid thermal process (RTP) is used as an industrial process to test the proposed unified framework after revising its PN-model. This scheme is not only developed for RTPs, but also can be employed for industrial processes with certain modifications depending on the nature of their structures. Simulation results have confirmed the effectiveness of the proposed unified framework.
Developing techniques for mobile robot navigation constitutes one of the major trends in the curr... more Developing techniques for mobile robot navigation constitutes one of the major trends in the current research on mobile robotics. This paper develops a local model network (LMN) for mobile robot navigation. The LMN represents the mobile robot by a set of locally valid submodels that are Multi-Layer Perceptrons (MLPs). Training these submodels employs Back Propagation (BP) algorithm. The paper proposes the fuzzy C-means (FCM) in this scheme to divide the input space to sub regions, and then a submodel (MLP) is identified to represent a particular region. The submodels then are combined in a unified structure. In run time phase, Radial Basis Functions (RBFs) are employed as windows for the activated submodels. This proposed structure overcomes the problem of changing operating regions of mobile robots. Read data are used in all experiments. Results for mobile robot navigation using the proposed LMN reflect the soundness of the proposed scheme.
This paper introduces a novel neural network named Auto Regressive eXogenous Local Model (ARX-LM)... more This paper introduces a novel neural network named Auto Regressive eXogenous Local Model (ARX-LM) network. It basically adopt the philosophy of forming a process be modeled with a set of fuzzy-wavelet submodels. Each submodel comprises a Takagi-Sugeno-Kang (TSK) fuzzy model fertilized by a wavelet function. The outputs of these submodels are weighted and summed to produce the final output. The main feature of the proposed LM network is that its structure is an ARX model that inherits the simplicity of the conventional linear and non-linear control systems, the learning capabilities of neural networks, the transparency of fuzzy systems and the locality of wavelet networks. The former means that the Lyapunove’s direct method can be employed to check the stability of the proposed ARX-LM network. The latter three features make the proposed network, powerful, transparent, and plastic. The plastic feature signifies that a new pattern can be assigned to uncommitted cluster if this pattern ...
Abstract: The Genetic Algorithm (GA) has the superiority over classical optimization algorithms i... more Abstract: The Genetic Algorithm (GA) has the superiority over classical optimization algorithms in finding the optimal (global) solution in multi-parameter search space. Because it simultaneously evaluates many points in the parameters space, it is more likely to converge toward the global solution. It need not assume that the search space differentiable or continuous and can also iterate several times on each datum received. This chapter proposes two genetic-based Takagi-Sugeno-Kang Recurrent Fuzzy Neural Network (TSK-RFNN) schemes; one is TSK-RFNN-G and the other is TSK-RFNN-GMRC. The former employs the well-known GA while the latter introduces the Model-Reference Control (MRC) to the former. The proposed algorithm uses real data instead of binary data to represent a chromosome. Unlike the TSK-RFNN-G, the TSK-RFNN-GMRC selects the controller’s parameters online (at each time step). For comparison reasons, both the proposed schemes were tested for controlling time intensive process...
Abstract: Elman network is a class of recurrent neural networks used for function approximation. ... more Abstract: Elman network is a class of recurrent neural networks used for function approximation. It has a set of global sigmoid functions at its hidden units. That means that if the operating conditions of a process be identified, are changed the function approximation property of the network is degraded. This is due to the fact that the universes of discourse of the network is covered by global sigmoid functions. This paper introduces a new version of the Elman network named Elman Recurrent Wavelet Neural Network (ERWNN). It merges the multi-resolution property of the wavelets and the learning capabilities of the Elman neural network to inherit the advantages of the two paradigms and to avoid their drawbacks. Structurally, the number of hidden/context units of an ERWNN should at least be equal to the order of the system be modeled. Stability and convergence property is proven for the proposed network. The dynamic back propagation (DBP) algorithm is employed to train the proposed ne...
Abstrncf - The performance of a model-based control system depends strongly on the accuracy of th... more Abstrncf - The performance of a model-based control system depends strongly on the accuracy of the process model used. This paper proposes a hybrid wavelet fuzzy neural network for modeling and control purposes. The proposed network inherits the advantages of both fuzzy and ...
The Genetic Algorithm (GA) has the superiority over classical optimization algorithms in finding ... more The Genetic Algorithm (GA) has the superiority over classical optimization algorithms in finding the optimal (global) solution in multi-parameter search space. Because it simultaneously evaluates many points in the parameters space, it is more likely to converge toward the global solution. It need not assume that the search space differentiable or continuous and can also iterate several times on each datum received. This chapter proposes two genetic-based Takagi-Sugeno-Kang Recurrent Fuzzy Neural Network (TSK-RFNN) schemes; one is TSK-RFNN-G and the other is TSK-RFNN-GMRC. The former employs the well-known GA while the latter introduces the ModelReference Control (MRC) to the former. The proposed algorithm uses real data instead of binary data to represent a chromosome. Unlike the TSK-RFNN-G, the TSK-RFNN-GMRC selects the controller’s parameters online (at each time step). For comparison reasons, both the proposed schemes were tested for controlling time intensive processes e. g. th...
Elman network is a class of recurrent neural networks used for function approximation. The main p... more Elman network is a class of recurrent neural networks used for function approximation. The main problem of this class is that its structure has a set of global sigmoid functions at its hidden layer. That means that if the operating conditions of a process be identified, are changed the function approximation property of the network is degraded. This paper introduces a new version of the Elman network named Elman Recurrent Wavelet Neural Network (ERWNN). It merges the multi- resolution property of the wavelets and the learning capabilities of the Elman neural network to inherit the advantages of the two paradigms and to avoid their drawbacks. Stability and convergence property is proven for the proposed network. The paper also develops a model reference control scheme using the proposed ERWNN. The proposed scheme belongs to indirect adaptive control schemes. The dynamic back propagation (DBP) algorithm is employed to train both the two networks structured for the indirect control sch...
Recently, an efficient method for obtaining reduced interpreted Petri net, IPN, diagnoser was pre... more Recently, an efficient method for obtaining reduced interpreted Petri net, IPN, diagnoser was previously proposed for discrete event systems. This diagnoser consists of a single place and the same number of transitions that the system model has. Although, the current marking of this place (diagnoser) is enough to determine and locate faults occurring in a simple diagnosable discrete event system, it fails to determine and locate faults occurring in a complex non-diagnosable hydride systems that have discrete and continuous variables. Diagnosability means that the PN model is live, strongly-connected, and Transition-invariant (T-invariant); however, most of PN models are not diagnosable. This paper generalizes the IPN-based diagnoser to deals with such complex systems with ease. Chemical batch processes are employed to test the modified IPN-based diagnoser and the simulation results obtained indicate that the proposed diagnoser is promising for industrial processes.
Elman network is a class of recurrent neural networks used for function approximation. The main p... more Elman network is a class of recurrent neural networks used for function approximation. The main problem of this class is that its structure has a set of global sigmoid functions at its hidden layer. That means that if the operating conditions of a process to be identified, are changed the function approximation property of the network is degraded. By noting the close analogy between neural networks, we may conclude that cross-fertilization between two different technologies may result a more powerful new one. Thus, this paper introduces a new version of the ELMAN network named ELMAN Recurrent Wavelet Neural Network (ERWNN). It merges the multi- resolution property of the wavelets and the learning capabilities of the ELMAN network to inherit the advantages of the two paradigms and to avoid their drawbacks. Stability and convergence property is proven for the proposed network. The dynamic back propagation (DBP) algorithm is derived to train the ERWNN. This paper also develops a model ...
2006 IEEE International Conference on Fuzzy Systems
Auto Regressive Moving Average (ARMA) models have attractive properties of simple computations an... more Auto Regressive Moving Average (ARMA) models have attractive properties of simple computations and fast learning. However, the unknown structures of the processes to be modeled produces difficult harmony between them and their ARMA models. This paper develops a novel ARMA model using the Particle Swarm Algorithm (PSA) to overcome this discord problem. It also introduces a PSA-based fuzzy control scheme
Uploads