Applied Mechanics and Materials Vols. 58-60

Abstract: To improve the control precision of multi-joint robots, a adaptive fuzzy immune PID control method for multi-joint robots was presented based on the immune feedback mechanism and fuzzy control theory, and the parameters of PID controller was optimized with hybrid algorithm. First, least square algorithm was used for off-line optimization to form immune feedback control system. Then, genetic algorithm was used for on-line optimization to get the optimal performance parameters of immune PID control system and the optimal fuzzy proportional parameters. Simulation results of a 2-joint robot manipulator demonstrated that the control method designed gets tracking effect with high precision and speed.

Abstract: A compressive sensing technique for image signal to cope with image compression and restoration is adopted in this paper. First of all wavelet transforms method is applied in image compressing to preserve the constructive, Secondly, sparse matrix is available by required wavelet ratio. Thirdly, the compressing image is used to restoration the original image. Experimental results show that the proposed algorithm is effective and compares favorably with existing techniques.

Abstract: In this paper, we proposed a punctuated target identification method for underground homogeneous medium based on dielectric constant inversion algorithm. Its main idea is that, initially, Hough transform is used to locate the target characterized by hyperbola in a radar profile map, then a layer-by-layer waveform inversion algorithm is used to invert the dielectric constant on the location where the target lies. To guarantee the correctness of inversion, the transmission beam method is adopted to obtain initial parameters and calibration factors required by inversion. Through numerical analysis of the dielectric constant of the target, the classification of multiple targets in an underground medium can be confirmed. This method not only overcomes the failure of the traditional phase-comparison method to distinguish different kinds of targets in the same region, but also overcomes the limitation of the image processing method, in which it only classifies the target in a coarse-grained manner. Experimental results show that this method has many advantages, such as fine-grained classification, high precision, and multiple target identification, in the identification of underground targets.

Abstract: The access strategy and method of control channel in WMN is an important research project. In this paper, description of access strategy and the method on distributed control channel in WMN based on TDMA are carried out, the coordination mechanism of multiple channels and allocation algorithm of controlling time slot to make the access of control channel fast and fair are proposed. Finally, the paper formulates protocol specification of control channel access, being appropriate for large-scale network, and discusses its application in the maintenance of network topology, route discovery and the establishment of QoS request.

Abstract: Aimed at the optimization problem of printed circuit board (PCB) assembly workshop (WPCBA) with multiple assembly machines and tasks, combined with polychromatic sets and genetic algorithm (GA), an optimization method is presented to optimize PCB assignment, component allocation and PCB assembly sequence problem simultaneously. On the basis of the polychromatic sets theory, numerical contour matrix is presented to describe the constraint of machine and operation sequence for WPCBA optimization problem and formulate the constraint model. Constraint model guarantees that GA searches optimal result in the effective solution space and simplify the calculation of fitness value. Moreover, if the machine and assembly task are changed, through a simple modification of constraint model, the WPCBA problem would be optimized conveniently. Computational results indicate that the solution efficiency of WPCBA optimization problem can be improved significantly and the dynamical optimization can be implemented.

Abstract: Identification is an important topic on simultaneous equation econometric model. It determines that whether the coefficients of the simultaneous equation models can be estimated. From the definition of identification perspective, this article will provide a new method about identification based on the relation between identified equations and the linear combinations of all equations.

Abstract: In the error detection mechanism of CAN bus, when the failed node or line is not in the data exchange path, you will not detect the fault, in order to solve this problem, a method is proposed which is realized by adding an external hardware detection circuit in the CAN-bus network, the fault will be detected through testing the resistance of the CAN bus network. In this paper, the network resistance model of CAN bus is established, and the principle of network resistance fault detection method is analyzed. We have carried out several experiments by the network with three nodes to test the validity of this method. As the results of our experiments, we concluded that the method can real-time and comprehensively detect the fault of network and do not take up the network bandwidth, so this method can effectively resolve the problems of current detection mechanism and have good application prospect in some high reliability requirements occasions.

Abstract: A novel image watermarking algorithm based on digital holography and ridgelet transform was designed by taking the advantage that ridgelet transform domain is a most suitable representation of image’s linear outline edge features. Firstly, the image was divided into blocks, and each block was transformed by finite ridgelet transformation. Secondly, the digital holography watermark that is got with the spread spectrum technique is embedded into the most energetic coefficients of ridgelet . A correlation function is used in the watermark detection process and the original image is not required.Experimen result show the watermark is robust to noise,medial filtering and JPEG2000 lost compression.By comparating respectively with paper[4] and paper[5],the result shows the proposed algorithm is more superior to them.

Abstract: The complexity of a real road network structure of a city and the variability of its real traffic information make a city’s intelligent transportation system (ITS) hard to meet the needs of the city’s vehicle navigation. This paper has proposed a simplified real-time road network model which can take into account the influence of intersection delay on the guidance for vehicles but avoid the calculation of intersection delay and troublesome collection of a city’s traffic data. Based on the new model, a navigation system has been presented, which can plan a dynamic optimal path for a vehicle according to the real-time traffic data received periodically from the city’s traffic center. A simulated experiment has been given. Compared with previous real-time road network models, the new model is much simpler and more effective on the calculation of vehicle navigation.

Abstract: This paper proposes a novel inner and outer eye corners detection method, which inosculates corner, regional texture and gray information, called CTGF algorithm. It utilizes corner detector to determine the candidate points of eye corners, such as Harris. Next, the region texture information is obtained through polar coordinate integral, in order to locate the exactly positions of eye corners among the candidate points. The CTGF algorithm provides a precise and reliable facial feature for many computer vision applications and the robustness and accuracy are demonstrated in experiments.