Applied Mechanics and Materials Vols. 433-435

Abstract: Center extraction of structured light’s stripe is a key link in the detection system for the robot’s passable areas based on a single-line structured light. With the original image that the camera took, we used the difference method to quickly get the interesting area where the light stripe was. Then we combined the Hough transformation with the improved K-means algorithm to take full advantage of the effective information of light stripe. We extracted the structured light’s center with high precision. It provides a guarantee for the robots’ passable area’s detection and path planning.

Abstract: In order to proactively provide value-added services for the electric vehicles, it proposes a three-tier grid-based collaborative design system framework. Dynamic and scalable wireless network based on grid collaboration can get mobile position of electric vehicles as well as exchange information among vehicles. Then it reduces the computational load on the backend server. By integrating power, transportation, municipal, and other areas of mass data meteorology, it can predict user needs based on characteristics of electric vehicles analytical model, and then provide users with real-time proactive value added services. This method avoids the presence of the GPS signal shielding and other shortcomings. The use of active value-added services better meets needs of customer.

Abstract: This paper introduced a new generation diagnosis system based on artificial neural network in diagnosing child Attention Deficit/ Hyperactivity Disorder (AD/HD). We give in detail the construction of the training set, including using computer to simulate thinking modes of human brain, to establish an artificial intelligence expert system for diagnosis and treatment of child AD/HD based on artificial neural network. The research can help the younger doctors to learn abundant clinical experiences of senior child psychiatrists and can help children of AD/HD throughout the country.

Abstract: The Back Propagation (BP) neural network was used for the construction of the hailstone classifier. Firstly, the database of the radar image feature was constructed. Through the image processing, the color, texture, shape and other dimensional features should be extracted and saved as the characteristic database to provide data support for the follow-up work. Secondly, Through the BP neural network, a machine for hail classifications can be built to achieve the hail samples auto-classification.

Abstract: Based on more than three hundred transformer accidents and some statistics data related, this paper manages to draw a standard classification table of transformer failure. 5 degrees of failure possibilities are classified in the paper: extremely high, high, general, low, extremely low. Besides, 4 major factors that affects risk of transformer failure severity are studied: direct lossmonitoring levelmaintenance cost and maintenance period. Also, failure severity of each risk source are classified into 5 levels: not serious, mild, ordinary, less serious, and serious, and Extensional Analytic Hierarchy Process (EAHP) and Extensional Project Appraisal Methods are used to evaluate the severity of risk of transformer failures. Last but not least, risk matrix has been built and divided into four parts: high risk area, critical risk area, medium risk area and low risk area. On that top, some maintenance strategies are proposed accordingly.

Abstract: As the lack in the accuracy and speed of the template matching algorithm for the snail image in the complex environment, the snail source image and the template image have the appropriate scaling in order to improve their sizes in the traditional algorithm. The new algorithm avoids the very big and accurate characteristics about the snail images through shrinking the source images down. The grayscale template matching method is put forward based on the traditional template selection set to prevent that the error caused by human factors on the selected template, the redundancy between the templates is removed in a large extent, further the accuracy of the matching is improved, and the matching time is reduced greatly in the case of matching accuracy guarantee.

Abstract: In fault diagnosis of three-phase induction motors, traditional methods usually fail because of the complex system of three-phase induction motors. Short circuit is a very common stator fault in all the faults of three-phase induction motors. Probabilistic neural network is a kind of artificial neural network which is widely used due to its fast training and simple structure. In this paper, the diagnosis method based on probabilistic neural network is proposed to deal with stator short circuits. First, the principle and structure of probabilistic neural network is studied in this paper. Second, the method of fault setting and fault feature extraction of three-phase induction motors is proposed on the basis of the fault diagnosis of stator short circuits. Then the establishment of the diagnosis model based on probabilistic neural network is illustrated with details. At last, training and simulation tests are done for the model. And simulation results show that this method is very practical with its high accuracy and fast speed.

Abstract: Neural network is acted as noise canceller to implement noise cancel under the condition of interference noise has nonlinear correlation to reference noise. If interference noise has nonlinear correlation to reference noise, the transversal filter has weak effect to cancel the noise in the signal. Neural network has nonlinear characteristic transfer and can solve this problem, and a new variable step size algorithm is proposed to further improve the performance. Computer simulation results show that neural network noise canceller has better signal to noise gain for nonlinear noise.

Abstract: Information service objects in agriculture relatively have a complex demand due to agricultural regional and seasonal. The construction of information service quality evaluation model contributes to analyze the influencing factors that influence the quality of information service, proving guidance for agricultural information service. Combined with genetic Algorithm, BP neural network and multiple regression, a hybrid BP network based on the integration of BP Network and multiple regression models is proposed, and the initial weights of hybrid BP network is optimized by hybrid genetic algorithm, effectively avoid the flaws when these methods used separately. Proved by the experiment, information service quality evaluation model constructed by a hybrid BP network based on the optimization of genetic Algorithm has a good accuracy and generalization ability, the mean error within 5%.

Abstract: The central issue of finishing train is that we should distribute the thickness of each exit with reason and determine the rolling force and relative convexity. The optimization methods currently used are empirical distribution method and the load curve method, but they both have drawbacks. To solve those problems we established a mathematical model of the finishing train and introduced an improved Genetic Algorithm. In this algorithm we used real number encoding, selection operator of a roulette and elitist selection and then improved crossover and mutation operators. The results show that the model and algorithm is feasible and could ensure the optimal effect and convergence speed. The products meet the production requirements.