Fuzzy Logic Control (FLC) is a promising control strategy in welding process control due to its ability for solving control problem with uncertainty as well as its independence on the analytical mathematics model. How...Fuzzy Logic Control (FLC) is a promising control strategy in welding process control due to its ability for solving control problem with uncertainty as well as its independence on the analytical mathematics model. However, in basic FLC, the fuzzy rule relies heavily on the experts’ (e.g. advanced welders’) experience. In addition to this, the membership function for fuzzy set is non adaptive, i.e. it remains unchanged as long as they are determined by experience or other means. For welding process, which is time variable systems and strong disturbance exists in it, fixed membership function may not guarantee the required system performance, and attempts should be made to improve the system performance by adopting adaptive membership function. Therefore, the automatically determination of the fuzzy rule and in process adaptation of membership function are required for the advanced welding process control. This paper discussed the possibility by using the combination between FLC and neural network (NN) to realize the above propose. The adaptation of membership function as well as the self organizing of fuzzy rule are realized by the self learning and competitiveness of the NN. Taking GTAW process welds bead width regulating system as the controlled plant, the proposed algorithm was testified for such a process. Computer simulations showed the improvement of the system characteristics.展开更多
A fuzzy observations-based radial basis function neural network （FORBFNN） is presented for modeling nonlinear systems in which the observations of response are imprecise but can be represented as fuzzy membership fu...A fuzzy observations-based radial basis function neural network （FORBFNN） is presented for modeling nonlinear systems in which the observations of response are imprecise but can be represented as fuzzy membership functions. In the FORBFNN model, the weight coefficients of nodes in the hidden layer are identified by using the fuzzy expectation-maximization （ EM ） algorithm, whereas the optimal number of these nodes as well as the centers and widths of radial basis functions are automatically constructed by using a data-driven method. Namely, the method starts with an initial node, and then a new node is added in a hidden layer according to some rules. This procedure is not terminated until the model meets the preset requirements. The method considers both the accuracy and complexity of the model. Numerical simulation results show that the modeling method is effective, and the established model has high prediction accuracy.展开更多
Fuzzy neural networks （FNN） based on Gaussian membership functions can effectively control the motion of underwater vehicles. However, their operating processes and training algorithms are complicated, placing great...Fuzzy neural networks （FNN） based on Gaussian membership functions can effectively control the motion of underwater vehicles. However, their operating processes and training algorithms are complicated, placing great demands on embedded hardware. This paper presents an advanced FNN with an S membership function matching the motion characteristics of mini underwater vehicles with wings. A leaming algorithm was then developed. Simulation results showed that the modified FNN is a simpler algorithm with faster calculations and improves responsiveness, compared with a Gaussian membership function-based FNN. It is applicable for mini underwater vehicles that don＇t need accurate positioning but must have good maneuverability.展开更多
文摘Fuzzy Logic Control (FLC) is a promising control strategy in welding process control due to its ability for solving control problem with uncertainty as well as its independence on the analytical mathematics model. However, in basic FLC, the fuzzy rule relies heavily on the experts’ (e.g. advanced welders’) experience. In addition to this, the membership function for fuzzy set is non adaptive, i.e. it remains unchanged as long as they are determined by experience or other means. For welding process, which is time variable systems and strong disturbance exists in it, fixed membership function may not guarantee the required system performance, and attempts should be made to improve the system performance by adopting adaptive membership function. Therefore, the automatically determination of the fuzzy rule and in process adaptation of membership function are required for the advanced welding process control. This paper discussed the possibility by using the combination between FLC and neural network (NN) to realize the above propose. The adaptation of membership function as well as the self organizing of fuzzy rule are realized by the self learning and competitiveness of the NN. Taking GTAW process welds bead width regulating system as the controlled plant, the proposed algorithm was testified for such a process. Computer simulations showed the improvement of the system characteristics.
基金The National Natural Science Foundation of China(No.51106025,51106027,51036002)Specialized Research Fund for the Doctoral Program of Higher Education(No.20130092110061)the Youth Foundation of Nanjing Institute of Technology(No.QKJA201303)
文摘A fuzzy observations-based radial basis function neural network （FORBFNN） is presented for modeling nonlinear systems in which the observations of response are imprecise but can be represented as fuzzy membership functions. In the FORBFNN model, the weight coefficients of nodes in the hidden layer are identified by using the fuzzy expectation-maximization （ EM ） algorithm, whereas the optimal number of these nodes as well as the centers and widths of radial basis functions are automatically constructed by using a data-driven method. Namely, the method starts with an initial node, and then a new node is added in a hidden layer according to some rules. This procedure is not terminated until the model meets the preset requirements. The method considers both the accuracy and complexity of the model. Numerical simulation results show that the modeling method is effective, and the established model has high prediction accuracy.
基金the Fundamental Research Foundation of Harbin Engineering University Foundation under Grant No.HEUFT08001
文摘Fuzzy neural networks （FNN） based on Gaussian membership functions can effectively control the motion of underwater vehicles. However, their operating processes and training algorithms are complicated, placing great demands on embedded hardware. This paper presents an advanced FNN with an S membership function matching the motion characteristics of mini underwater vehicles with wings. A leaming algorithm was then developed. Simulation results showed that the modified FNN is a simpler algorithm with faster calculations and improves responsiveness, compared with a Gaussian membership function-based FNN. It is applicable for mini underwater vehicles that don＇t need accurate positioning but must have good maneuverability.
