Papers by Keyword: Fuzzy Neural Network (FNN)

Abstract: To reduce false fire alarms, combining with the character of fire signal, a kind of intelligent fire detection system of multi-sensor information fusion based on fuzzy neural network is proposed in this paper . This fire detector fuses three sensor data including temperature, smoke and CO air which have obvious character in fire and fire probability can be obtained by intelligent arithmetic of fuzzy neural network. As a result, The accuracy of the fire detection is improved effectively and the feasibility and validity of the system are proved by the simulation effects. 0 Foreword The purpose of fire detection technology is to make accurate judgments of the fire and to predict the fire in the early time, so that people's lives and property can be protected. Based on the monitoring of physical phenomena such as light, smoke, heat, the traditional fire detection usually monitors one kind of physical quantity and establishes a certain threshold value as the criterion for the fire. In practice, it is discovered that fire monitoring, based on a certain physical quantity and threshold value, is often inevitably influenced by a certain similar environmental factors influence which causes false alarm. 1 Multi-sensor Data Fusion Fire Detection System For any kind of detective object, using only one kind of information to reflect its condition is not complete. Only through getting, integrating and using various multi-dimensional information of the same object, it can detect the fire accurately and early. In view of the fact that unit fire detection technology has been unable to meet the needs of real fire alarm, the system uses multiple information fusion fire detection, which is not the simple combination of the fire detectors original single parameter, but the implementation of multiple simultaneous detection, extraction of useful and accurate information. According to different types of fire parameters, it applies intelligent algorithms, fuses the fire parameters of multi-sensor fusion, and determines whether there is a fire hazard. It overcomes the limitations of a single sensor, and effectively improves the ability of identifying real or false fires. Under normal circumstances, CO is extremely low in the air. Only by burning massive CO can be produced, which causes the density of CO in the air to increase sharply. Thus the detection of CO gas will be in large part reflects whether the combustion phenomenon happens or not. The occurring of fire is often accompanied with the elevation of temperature and the enlargement of smoke density, so the system of fire detectors uses 3-layer structure of multi-sensor fusion, selects temperature sensors, smoke sensors, gas sensors, the temperature signal, smoke concentration and the CO concentration as the fire detection signal. 2 Fuzzy Neural Network Applying fuzzy neural network to fire detection information processing can greatly improve the timeliness and accuracy of fire detection, and reduce the rate of false alarm.This system uses fuzzy neural network as shown in Figure 1. Before and after the neural network in the system is in series with the fuzzy system, in order to facilitate the procession of neural network, the smog density signal from the environment examination, the temperature signal as well as the gas signal through the signal pretreatment should be normalized, and sends these three normalized values into the fuzzy system, uses trigonometric functions for transformation, and obtains three degree of membership and the feedback signal of neural network as the neural network input.

Abstract: In order to realize the precise identification of eggshell crack, we design eggshell cracks detection method based on image processing and fuzzy neural network. Firstly this method gets two pieces image of eggs and processes, and then counts number of the same gray pixel. Determine five characteristic parameters as the input of fuzzy neural network. Set up a fuzzy neural network. Its structure is 5-10-1. Eggshell cracks and noise in egg images were distinguished using automatically learning and inference rules of fuzzy neural network. Use 147 groups of parameters for training network and rest 58 sample for verifying. Experimental result shows that the model can meet actual testing requirements with fast, stable, high precision and good robustness, easy to implement. Its precision reached 94.55%.

Abstract: In this article, a new combustion control program would be introduced, which is planned to use fuzzy neural network for intelligent analysis of temperature variation on end fire glass furnace. This program would be adapted to regenerative end fire glass furnace. The actual running of system indicates that the system is in good effect and suitable for the production technology very well.

Abstract: Multi-sensor information fusion is that fuses information of multiple sensors gained through use of redundant, complementary, or timelier information in a system can provide more reliable and accurate information. Under the research of mobile robot environmental information, a control method of fuzzy neural network based on T-S (Takagi-Sugeno) type is given, it can fuses effectively collected information from multiple ultrasonic sensors and a CCD camera, and realize the real-time control for mobile robot. The results on mobile robot obstacles avoidance verified the effectiveness of the method.

Abstract: This paper presents a hybrid GA-BP algorithm for fuzzy neural network controller (FNNC). BP algorithm is a method to monitor learning, easily realized and with good local searching ability. But it depends too much on the the initial states of the network. Genetic algorithm is a random search algorithm which has strong global searching ability. The hybrid GA-BP algorithm ensure the global convergence of learning by genetic algorithm, overcomes the BP algorithms dependency on the initial states on the one hand. On the other hand, combined with the BP algorithm overcome the simple genetic algorithms randomness, improve the searching efficiency. The simulations on the inverted pendulun problem show good performance and robustness of the proposed fuzzy neural network controller based on hybrid GA-BP algorithm.

Abstract: Using fuzzy neural network to tune PID parameters, and DSP as processor, it was designed that a set of electric boiler temperature control system based on PID parameters self-tuning, including the design of each hardware module and each software subroutine of the system. Experimental results show that compared with the traditional PID temperature control system, this temperature control system has the advantages such as good control effect, easy parameter adjustment, strong anti-jamming capability, better adaptability and robustness, has the feasibility and practical value.

Abstract: The correct inventory strategy is important to control reasonable inventory within manufacturing enterprises. And a new material inventory strategy is proposed based on the classification scheme according to the prosperities. First the material inventory styles are classified. Then a decision tree model is defined based on inventory classification result. The value of the node is decided by Fuzzy Neural Network if multi-attribute decision is needed and material inventory strategy can be decided with the classification tree and inventory strategy table. In the end, the implementation of the model in a manufacturing enterprise resource plan system is presented.

Abstract: In order to improve egg hatchability with non-destructive identification of fertilized eggs, a system that captures egg image online and extracts image feature parameters is designed. It takes an TMS320DM642 processing chip as the core and sets up a fuzzy neural network which is based on 5-D input and 1-D output to identify the egg fertilization. The test result indicated that this model has a high speed and accuracy, easy to realize with 97.22% accuracy as matching the requirement of actual test.

Abstract: Firstly, the implication and purpose of the Combat Service Performance Evaluation for Air Defense Missile was elaborated in this paper. Then the Fuzzy Neural Network (FNN) Evaluating Method was put forward to apply to the combat service performance evaluation. Secondly, an evaluation index system was built on the basis of the characteristics of the combat service operator and the combat service process. At last, the correctness and validity were approved though the simulation results.

Abstract: This Paper focuses on the semi-active control application in the Mega-Sub Controlled Structure System (MSCSS) subjected to seismic excitation. The semi-active control devices, which are installed in the MSCSS between the mega-structure and sub-structure, were designed by using fuzzy neural network, and those semi-active control rules were optimized to enhance the control efficiency by using the genetic algorithm. A semi-active control problem of the MSCSS subjected to seismic excitation was investigated, the time history analyses under different seismic excitation, which like El Centro seismic wave and Taft seismic wave, were performed by using MATLAB. The calculation results demonstrate that the semi-active control combining the fuzzy neural network and genetic algorithm can clearly enhances the seismic performance of the MSCSS and it also provides an improved reduction in the dynamic response when compared to the passive control scheme.