Applied Mechanics and Materials Vols. 121-126

Abstract: In this paper, an approach based on local curvature feature matching for 3D face recognition is proposed. K-L transformation is employed to adjust coordinate system and coarsely align 3D point cloud. Based on B-splines approximation, 3D facial surface reconstruction is implemented. Through analyzing curvature features of the fitted surface, local rigid facial patches are extracted. According to the extracted local patches, feature vectors are constructed to execute final recognition. Experimental results demonstrate high performance of the presented method and also show that the method is fairly effective for 3D face recognition.

Abstract: During the development of the facial expression classification procedure, we evaluate three machine learning methods. We combine ABAs with CARTs, which selects weak classifiers and integrates them into a strong classifier automatically. We have presented a highly automatic facial expression recognition system in which a face detection procedure is first able to detect and locate human faces in image sequences acquired in real environments. We need not label or choose characteristic blocks in advance. In the face detection procedure, some geometrical properties are applied to eliminate the skin color regions that do not belong to human faces. In the facial feature extraction procedure, we only perform both the binarization and edge detection operations on the proper ranges of eyes, mouth, and eyebrows to obtain the 16 landmarks of a human face to further produce 16 characteristic distances which represent a kind of expressions. We realize a facial expression classification procedure by employing an ABA to recognize six kinds of expressions. The performance of the system is very satisfactory; whose recognition rate achieves more than 90%.

Abstract: Toothbrush is a simple but an indispensible product. The reasonable or not of its design directly influences human’s oral health. Its handle is the main body of a toothbrush and plays a critical role in the success of its design. This paper makes a research on toothbrushes and studies on the general design principle of toothbrush handle with a combination of ergonomic theory.

Abstract: A method to get numerical solutions to Nonlinear Cyber-Security Equationss with the the N algorithm is proposed . This paper especially focused on its converging rate and application. Also, It did some comparison between with other universal algorithms. Both mathematical method and experimental method are used to do the comparison. This paper gave strict mathematical proof and also used the results we got from our program to draw a solid conclusion.

Abstract: In order to solve problems concerning the complex module replacement and positioning difficulties in the modular design process, the control-model which reflects the main structure of product is established. Control-model bases itself on the product’s absolute coordinate system, and by establishing datum plane and sketches it reflects the main structure of products. According to the relationship between modules, modules can be partitioned into main-modules and attached-modules. The main- and attached-modules established by using control-model as seed file contain relevant information of the control-model, therefore the method of modules’ rapid positioning by comparison and updating as well as module change strategies can be used to realize modular design process without assembly constraints. Finally, taking a certain type of tank semi-trailer as an example, this paper makes detailed explanation about the modular design method based on control-model.

Abstract: A design synthesis technique based on sensitivity for Micro-Electro-Mechanical Systems (MEMS) proposed. This new technique can be called Sensitivity-Based Direct Solution Algorithm (DSA) of design synthesis for MEMS with expected performance. Design synthesis with expected performance is regarded as a reverse problem of MEMS analysis. Behavior equation group can be educed from analysis equations. This behavior equation group can be solved using any solution algorithm of non-linear equation group. Newton Iteration Method based on sensitivity is adopted. Comparing with Genetic Optimization Algorithm (GA) , computational workload of DSA is greatly decreased.

Abstract: Plasma Chemical Vapor Deposition ( PCVD ) is an important application in making optical fiber core rod. The resonator is the key part for PCVD application. But the existing resonator have several serious disadvantages: its working is not stable; the size of deposited substrate tube is small(Ø31mm); the deposition rate is low(2.5g/min) and the deposition effect is unsymmetrical; the fabrication of the resonator is very difficult and expensive. Therefore, a new type of resonator, a cylinder resonator, is designed to overcome the above disadvantages. The cylinder resonator supports bigger substrate tube (Max Ø60mm) and higher deposition rate(3.2g/min) than existing ones, while its working is also more stable. Since the structure is more simple, the fabrication is easier and cheaper. The corresponding microwave theoretical method of the new resonator designing is also introduced with the necessary computer simulation, as well as test running. And TE111 mode microwave is the Fundamental Mode of the cylinder resonator. The cylinder resonator has been applied in normal production for 4 years, which is better for bigger size core rod making, improving the optical fiber quality and saving cost. It is estimated, after modification with the cylinder resonator and 10kW microwave generator, a PCVD lathe can yield 30% more rods than before.

Abstract: This paper investigates the development and intelligent modeling problem for a wood drying kiln process via optimized support vector machine (SVM). Based on parameters optimization and model selection idea, the swarm intelligence algorithms of Particle Swarm Optimization (PSO)-SVM and Genetic Algorithm (GA)-SVM were proposed for wood drying process with strong coupling and nonlinear characteristics. The simulation results showed that both of these two kinds of swarm intelligence optimization algorithm could get the appropriate parameters of SVM effectively, and by contrast, PSO showed a better learning ability and generalization in wood drying process modeling, and could establish predictive model with better accessibility.

Abstract: This research proposes a novel low temperature manufacturing method to make a wireless accelerometer on a flexible substrate. The substrate deposition temperature is 100°C without causing any strain and stress problem. Since the thermal conductivity of the traditional Si is 1.48 W/ (cm-K), which is 25 times of the flexible substrate, i.e. 0.06-0.0017 W/ (cm-K), thus the power leakage through the substrate can be saved by the new design. The key technology is to integrate a thermal bubble accelerometer and a wireless RFID antenna on the same substrate, such that the accelerometer is very convenient for fabrication and usage. In this paper the heaters and the thermal piles are directly adhering on the substrate surface without the traditional floating structure. Thus the structure is much simpler and cheaper for manufacturing, and much more reliable in large acceleration impact condition without broken. Furthermore, the molecular weight of xenon gas (131.29 g/mol) is much larger than carbon dioxide (44.01 g/mol), thus the performance of the accelerometer will be increased. In addition, the shape of the chamber is changed as a semi-cylindrical one instead of the conventional rectangular type. The average sensitivity is increased by 15%. In addition, if one applies only xenon gas but keeping the rectangular chamber, then the response speed can be increased by 23%. Moreover, if one applies both Xe and the semi-cylindrical chamber, then the response speed can be increased by 43%.

Abstract: This paper combines Genetic Algorithm with Simulated Annealing Algorithm, namely GA-SA，to discuss vehicle paths and take into account the condition of time with respect to multi-spot service combination problem in service centre. The prevalent genetic algorithms easily lose the optimal solution, which affects the entire algorithm performance for reality vehicle assignment problem in the service centre. Based on modelling the vehicle assignment problem with natural description, fitness function, crossover operation and mutation operation are made the improvement in the approach. The process of computation has also considered own characteristics of the service centre to enable the algorithm optimized performance, in order to obtain the large scale enhancement.