Advanced Materials Research Vols. 542-543

Abstract: To locate pedestrian faster and more accurately, a pedestrian detection method based on histograms of oriented gradients (HOG) in region of interest (ROI) is introduced. The features are extracted in the ROI where the pedestrian's legs may exist, which is helpful to decrease the dimension of feature vector and simplify the calculation. Then the vertical edge symmetry of pedestrian's legs is fused to confirm the detection. Experimental results indicate that this method can achieve an ideal accuracy with lower process time compared to traditional method.

Abstract: By analyzing the structure features and tool changing process of a large-scale national Automatic Tool Changer (ATC) with a chain-type tool magazine, a chain-type tool magazine reliability test bench is designed with a new fault detection system, where a new failure data collecting method is applied for taking failure data statistics and reliability analysis. This set of fault detecting and reliability analyzing system not only provides a way for ATC to predict system faults, but also creates good conditions for further ATC reliability research.

Abstract: The paper studies the microspheres etching technique, the silicon pillar arrays are fabricated using polystyrene particles as etching mask by Reactive Ion Etching. Obtained the substrate can be Lateral epitaxial. The influence of different parameters in etch process are investigated on silicon pillar arrays in detail. A large-area Si pillar can be obtained on Si substrate by controlling the suitable etch parameters. The approach reported here offers a possibility to product large-area pillar.

Abstract: According to the required functions of speed control and position control in Computer Numerical Control (CNC) system, the hardware control modules of speed and position are designed based on Field Programmable Gates Array (FPGA) of CYCLONE II family. The software hardening technology is used for speed control and position control. The servo intelligent control chip consists of reset unit, frequency division unit, speed processing unit, subdivision unit, phase discrimination unit, counter unit, compare unit, etc. Through analysis of waveform simulation, the correctness of design and the enhancement of transportability and reliability are achieved.

Abstract: An Al_0.3Ga_0.7As/GaAs/Al_0.3Ga_0.7As/GaAs/Al_0.3Ga_0.7As asymmetric dual-quantum-well structure was designed. The barrier thickness between the dual quantum wells is 48Å, it is thick enough to inhibit the mutual cross-interference between the energy levels within the two quantum wells. This material system was grown on a GaAs substrate by solid source molecular beam epitaxy, and the device was fabricated with rat electrodes using inductively coupled plasma etching mesa process. The tunneling effect that electron transfer through the dual-quantum-well structure was observed in the device I-V feature, it was calculated and demonstrated by transmission matrix method and Ariy Function numerical transform.

Abstract: A quartz crystal microbalance (QCM) modified with a thin layer of fluorinated polymer was utilized to determine dimethylmethyl phosphonate (DMMP). Determination was based on the frequency shifts due to the adsorption and desorption of the compound at the surface of modified quartz crystal electrode. Fluorosiloxane was utilized onto the surface of the sensor. The frequency shifts were linear against the concentration of DMMP within the concentration range of 30~210mg•m^-3. Good selectivity and sensitivity were obtained with the sensor for DMMP. The detection limit was 10.1 mg•m^-3.

Abstract: The multiple degrees of freedom mechanical arm of sampling machine for coal at railway carriage is driven by hydraulic press. It is a kind of joint. During working, the sampling head not only keeps in its poses, but also moves along a straight path. The mechanical arm’s motion is resolved into three independent motion systems, foundation rotation, straight motion of 3 degrees of freedom kinematics chain, and servo control system of the sampling head. A group of motion equations is set up. According to the reversion theory and optimization method, the motion control model is built up in order to reappear the motion track. The method is efficient to solve the motion control problem. And the result of driving curve is more suitable for the hydraulic system. This method is useful for motion intelligent control of different kinds of joint manipulator.

Abstract: Combined with the knowledge of BP neural network and template matching, a new face recognition method was given. It matched the face image which was identified with the serial number and corresponding to the face model that based on random sequence in the matching process. The acquisition image should be preprocessed in filtering, smoothing and sharpening way by using BP neural network learning ability and classification ability which can achieve classifier. By using the actual face, it verify the correctness and reliability of the hardware and software design, manifest a good authentication capabilities and provide a reference for the research of similar products.

Abstract: To analyze the intelligent measurement and control system in real world theoretically and overcome the disadvantages, where the high-general and abstract models are lacking for the analysis, guidance and description, a structure model was proposed based on personification theory. In this model, several major functional units of system were extracted using human body system as the analysis prototype. And three aspects of U(structural unit set), S(service set) and R(the relationship set) were discussed. The theory analysis suggests that this model has general, abstract, analyzable and guidable characteristic for analyzing intelligent measurement and control system.

Abstract: In this article, the parameters affecting the quality of wire bonding are analyzed by orthogonal testing with the methods of variance analysis and F tests. By analyzing the results, parameters that have a major impact on the quality of wire bonding are optimized. Because the relationship is complicated and non-linear between the impacting parameters and bonding quality, this article introduces a neural network algorithm of BPNN to build a model describing it. The structural parameters of the neural network are identified and a quality prediction model of wire bonding is established in this article. The model is validated, the results show that this proposed model has higher precision and it can accurately reflect the trends of the bonding quality indicators.