Applied Mechanics and Materials Vols. 513-517

Abstract: Real-time image processing has been a difficult problem in embedded image processing system. The traditional MCU could not meet the real-time demand when large volume of data awaited to be proceed. FPGA is an effective driver to achieve real-time parallel processing of data. The implementation rationale and the design of module have been given in this article; and the Hard Software has been truly achieved. At the end of the article, the simulation waveform graph has been obtained by processing functional simulation on algorithm module by using Modelsim software; and the simulation result shows that this design is able to proper functioning and has good application prospects.

Abstract: To improve the ability of undertaking large projects in colleges, this paper put forward a detailed development of College Laboratory Construction Project Management System on the analysis and construction of laboratory. This management system was developed with myApps which is highly useful for developing a management system based on workflow.

Abstract: In the view of frequency excursion estimation in cell search process in TD-LTE system, synchronizing signal frequency excursion estimation algorithm can be proposed based on the autocorrelation of synchronizing signal. Integer frequency excursion and fine frequency excursion can be estimated according to the correlating peak value, so that frequency offset can be made. The simulation results show that frequency excursion can be estimated exactly by the synchronizing signal.

Abstract: In order to solve the limitations of classical packet scheduling architecture in TD-LTE, the TD-LTE downlink packet scheduling was studied and a architecture of downlink packet scheduling for TD-LTE is designed. The packet scheduling architecture was built and three classic scheduling algorithms were simulated in this architecture by Matlab. The simulation results show that the packet scheduling architecture can be applied to scheduling algorithms with different purposes and verify the performances of throughput and fairness of these three scheduling algorithms.

Abstract: Identification and modification of real-time traffic data has been the basic and critical part in the intelligent transportation system.Through the research to a large number of data,the original data is divided into the correct data,the irregular time-point data,inaccurate detection data,missing data and event data. Etkin interpolation algorithm is to gain the values of specified missing value by a successive approximation method with high order polynomial and implemented by using a successive approximation of multiple linear combinations.The paper selects improved Etkin interpolation algorithm to correct the traffic data and makes use of the DongZhiMen Bridge North 728 meters' 2001 detector data for example.The algorithm not only considers the practicability in the engineering practice,but also improves the accuracy of real-time data.

Abstract: According to analysis of manufacturing enterprise coding requirements, Visual flexible coding system (VFCS) is proposed based on code classification theory. It has four function modules: code structure design, code segment & bit design, visual coding and interpreting, code query. Firstly, code rules are expressed by extendable XML format as data structure. Secondly, code design is regulated into two stages: code structure design and code segment design, so as to realize separation of code structure logical and contents. finally, code segment structure information is described as graphic nodes, using interactive graphics drawing for visual modeling, finite state machine algorithm are used implementing code structure interpreter.

Abstract: The collection and management of dynamic traffic information is one of the most important part of ITS. Its a main task for it to improve the accuracy of the acquisition of the traffic information when facing up with different kinds of traffic detectors. Data fusion method can deal with data from different detectors and improve the accuracy. This paper first analyzed the characters of different traffic detectors, and proposed a method to repair the missing values which is a common phenomenon in the detect data. Then some improvements are made to adjust the BP neural network so that it could be suitable for data fusion. At last, the data fusion of traffic speed from the south of Jianguomen Qiao to the north of Chaoyangmen Qiao on the second ring road of Beijing is given as an example with the comparison of different improve methods of BP neural networks, and it shows that the method given in this passage is efficient in improving the accuracy of the traffic data detection.

Abstract: Partially Observable Markov Decision Processes (POMDP) provides piecewise-linear a natural and principled framework for sequential decision-making under uncertainty. However, large-scale POMDP suffers from the exponential growth of the belief points and policy trees space. We present a new point-based incremental pruning algorithm based on the piecewise linearity and convexity of the value function. Instead of reasoning about the whole belief space when pruning the cross-sums in POMDP policy construction, our algorithm uses belief points to perform approximate pruning by generating policy trees, and get the optimal policy in real-time belief states. The empirical results indicate that point-based incremental pruning for heuristic search methods can handle large POMDP domains efficiently.

Abstract: Bayesian reinforcement learning has turned out to be an effective solution to the optimal tradeoff between exploration and exploitation. However, in practical applications, the learning parameters with exponential growth are the main impediment for online planning and learning. To overcome this problem, we bring factored representations, model-based learning, and Bayesian reinforcement learning together in a new approach. Firstly, we exploit a factored representation to describe the states to reduce the size of learning parameters, and adopt Bayesian inference method to learn the unknown structure and parameters simultaneously. Then, we use an online point-based value iteration algorithm to plan and learn. The experimental results show that the proposed approach is an effective way for improving the learning efficiency in large-scale state spaces.

Abstract: In order to improve the neural network structure and setting method of parameters, based on the glowworm swarm optimization (GSO) and BP neural network (BPNN), an algorithm of BP neural network optimized glowworm swarm optimization (GSOBPNN) is proposed. In the algorithm, GSO is used to obtain better network initial threshold and weight so as to compensate the defect of connection weight and thresholds choosing of BPNN, thus BPNN can have faster convergence and greater learning ability. The efficiency of the proposed prediction method is tested by the simulation of the chaotic time series of tent mapping. The simulations results show that the proposed method has higher forecasting accuracy compared with the BPNN, so it is proved that the algorithm is feasible and effective in the chaotic time series.