In a special case of type-2 fuzzy logic systems （FLS）, i.e. geometric inteIval type-2 fuzzy logic systems （GIT-2FLS）, the crisp output is obtained by computing the geometric center of footprint of uncertainly （FO...In a special case of type-2 fuzzy logic systems （FLS）, i.e. geometric inteIval type-2 fuzzy logic systems （GIT-2FLS）, the crisp output is obtained by computing the geometric center of footprint of uncertainly （FOU） without type-reduction, but the defuzzifying method acts against the corner concepts of type-2 fuzzy sets in some cases. In this paper, a PSO type-reduction method for GIT-2FLS based on the particle swarm optimization （PSO） algorithm is presented. With the PSO type-reduction, the inference principle of geometric interval FLS operating on the continuous domain is consistent with that of traditional interval type-2 FLS operating on the discrete domain. With comparative experiments, it is proved that the PSO type-reduction exhibits good performance, and is a satisfactory complement for the theory of GIT-2FLS.展开更多
Adaptive Type-2 fuzzy control possesses control performance better than the traditional adaptive fuzzy control.However,heavy computation burden obviously blocks the utilization of adaptive Type-2 fuzzy control in indu...Adaptive Type-2 fuzzy control possesses control performance better than the traditional adaptive fuzzy control.However,heavy computation burden obviously blocks the utilization of adaptive Type-2 fuzzy control in industrial application.By adopting novel piecewise fuzzy sets and center-average type-reduction,a simplified adaptive interval Type-2 fuzzy controller involving less computation is developed for practical industrial application.In the proposed controller,the inputs are divided into several subintervals and then two piecewise fuzzy sets are used for each subinterval.With the manner of piecewise fuzzy sets and a novel fuzzy rules inference engine,only part of fuzzy rules are simultaneously activated in one control loop,which exponentially decreases the computation and makes the controller appropriate in industrial application.The simulation and experimental study,involving the popular magnetic levitation platform,shows the predicted system with theoretical stability and good tracking performance.The analysis indicates that there is far less computation of the proposed controller than the traditional adaptive interval Type-2 fuzzy controller,especially when the number of fuzzy rules and fuzzy sets is large,and the controller still maintains good control performance as the traditional one.展开更多
This paper presents an analysis of the KM （Karnik-Mendel） algorithms performance under real time implementation using 3 types： the non-iterative, the iterative and the enhanced, and their feasibility for real-time ...This paper presents an analysis of the KM （Karnik-Mendel） algorithms performance under real time implementation using 3 types： the non-iterative, the iterative and the enhanced, and their feasibility for real-time interval type 2 fuzzy logic control system applications. The results are also compared against NT （Nie-Tan） method that is one of the fastest and simplest defuzzification methods. Because the DC （direct current） servo-motor is one of the most used motors in different industrial applications and the model of the motor is nonlinear, this motor was selected for validating the implementation in real time hardware. This DC motor is a perfect option for studying the real time performance of KM algorithms in order to show up its limits and possibilities for real-time control system applications. These methodologies are implemented in National Instruments LabVIEW FPGA （field programmable gate array） module hardware which is one of the most used platforms in the industry. The results show that the E-KM （enhanced KM） algorithm and the NT method present good results for implementing real-time control applications in real time hardware. Although fuzzy logic type 2 is a good option for working with nonlinear and noise from the sensors, the defuzzification method has to react in a short period of time in order to allow good control response. Hence, a complete study of defuzzification is needed for improving the real time implementations of fuzzy type 2.展开更多
基金Sponsored by the National Hi-Tech Program of China(Grant No. 2005AA420050)the National Key Technology R&D Program of China(Grant No.2006BAD10A0401, 2006BAH02A01)
文摘In a special case of type-2 fuzzy logic systems （FLS）, i.e. geometric inteIval type-2 fuzzy logic systems （GIT-2FLS）, the crisp output is obtained by computing the geometric center of footprint of uncertainly （FOU） without type-reduction, but the defuzzifying method acts against the corner concepts of type-2 fuzzy sets in some cases. In this paper, a PSO type-reduction method for GIT-2FLS based on the particle swarm optimization （PSO） algorithm is presented. With the PSO type-reduction, the inference principle of geometric interval FLS operating on the continuous domain is consistent with that of traditional interval type-2 FLS operating on the discrete domain. With comparative experiments, it is proved that the PSO type-reduction exhibits good performance, and is a satisfactory complement for the theory of GIT-2FLS.
基金Project(51005253) supported by the National Natural Science Foundation of ChinaProject(2012ZX02702006-003) supported by the National Science and Technology Major Program of ChinaProject(JMTZ201101) supported by the Key Laboratory for Precision & Non-traditional Machining of Ministry of Education,Dalian University of Technology,China
文摘Adaptive Type-2 fuzzy control possesses control performance better than the traditional adaptive fuzzy control.However,heavy computation burden obviously blocks the utilization of adaptive Type-2 fuzzy control in industrial application.By adopting novel piecewise fuzzy sets and center-average type-reduction,a simplified adaptive interval Type-2 fuzzy controller involving less computation is developed for practical industrial application.In the proposed controller,the inputs are divided into several subintervals and then two piecewise fuzzy sets are used for each subinterval.With the manner of piecewise fuzzy sets and a novel fuzzy rules inference engine,only part of fuzzy rules are simultaneously activated in one control loop,which exponentially decreases the computation and makes the controller appropriate in industrial application.The simulation and experimental study,involving the popular magnetic levitation platform,shows the predicted system with theoretical stability and good tracking performance.The analysis indicates that there is far less computation of the proposed controller than the traditional adaptive interval Type-2 fuzzy controller,especially when the number of fuzzy rules and fuzzy sets is large,and the controller still maintains good control performance as the traditional one.
文摘This paper presents an analysis of the KM （Karnik-Mendel） algorithms performance under real time implementation using 3 types： the non-iterative, the iterative and the enhanced, and their feasibility for real-time interval type 2 fuzzy logic control system applications. The results are also compared against NT （Nie-Tan） method that is one of the fastest and simplest defuzzification methods. Because the DC （direct current） servo-motor is one of the most used motors in different industrial applications and the model of the motor is nonlinear, this motor was selected for validating the implementation in real time hardware. This DC motor is a perfect option for studying the real time performance of KM algorithms in order to show up its limits and possibilities for real-time control system applications. These methodologies are implemented in National Instruments LabVIEW FPGA （field programmable gate array） module hardware which is one of the most used platforms in the industry. The results show that the E-KM （enhanced KM） algorithm and the NT method present good results for implementing real-time control applications in real time hardware. Although fuzzy logic type 2 is a good option for working with nonlinear and noise from the sensors, the defuzzification method has to react in a short period of time in order to allow good control response. Hence, a complete study of defuzzification is needed for improving the real time implementations of fuzzy type 2.
