Search

Marine environmental data field is a kind of high-dimensional and multi variable spatial-temporal data field, which is large-volume, time-varying, multi-format and has strong diversity. Spatiotemporal data is the basis of data browsing and data analysis, so improving the efficiency of retrieval has become a reliable solution to improve the timeliness of processing. Aiming at the organization and spatial retrieval of massive marine spatiotemporal data. Our work was carried out from the following aspects, one is to unify the storage organization and build a data model for ocean spatiotemporal data, the other is to construct distributed spatial data index of ocean spatiotemporal data. Ocean spatiotemporal data model realizes the unified organization modeling of data with different sources, high dimensions, diverse forms, heterogeneous formats and high spatial temporal connection. We describe this model from data sources, collection time and geospatial information. Based on the strategy of data segmentation and spatial region division, two kinds of distributed spatial data retrieval indexes are designed and implemented: single level R-tree spatial distributed data index based on ocean spatiotemporal data fragmentation and multi-level spatial distributed data indexing of R-tree + grid index based on global geospatial partitioning. The experiment results prove that our strategy could effectively minimize the interference of irrelevant data when retrieving, reduce retrieval redundancy, efficiency enhanced the task timeliness of the ocean spatiotemporal data processing platform.

In this paper, summarizing the problems in the operation and maintenance, and based on data syncretism, research which is on the outdoor installation protection of multi-source perception and intelligent diagnosis technology, designs an outdoor installation protection perception and operational maintenance platform that highly integrated protection on-line monitoring and diagnosis, protection information sub-station, power system dynamic recording equipment and network message recording and analysis device as one system, and expounds in detail the system architecture of the platform, software framework, realization principle and key technology. The platform through the multi-source data acquisition, and the data sharing between each module and system integration, implements the outdoor installation protection multi-source perception and intelligent diagnosis, improving the efficiency and convenience of field operation and maintenance work, for outdoor installation protection of the safe and stable operation is of great significance. The platform has been successfully carried out in substation pilot application, achieving good results.

In order to improve the ability of the officers to control the complex electromagnetic environment on the battlefield, a method of calculating and constructing the electromagnetic environment by means of the digital earth, the large data of the spectrum and the situation simulation technology, the radiation source and the wave propagation model is presented. Realize the integrated electromagnetic situation visualization.

Hydraulic support optimization can improve the capabilities of managing the roof and enduring the load of the standing shield's hydraulic support. In this paper, an improved particle swarm optimization (IPSO) algorithm is proposed to optimize the hydraulic support. In order to avoid the particles falling into local searching and improve the particles convergence rate, the strategy of updating the particle's velocity and location is improved, and a new nonlinearly decreasing function is constructed as inertia weight. Taking the length of four-bar linkage of hydraulic support as the optimal parameters and the maximal horizontal swing ¦¤x of the move trace of the roof beam endpoint minimum as the optimal target, we build parameter optimization model of four-bar linkage based on IPSO and make comparison with analytical method. The results show that, the ¦¤x is greatly reduced; the working status and support effect of hydraulic support is stable and favorable.

The essay is to apply the “Cloud Manufacturing” to the mould industry via the analysis on the mould design and manufacture of mould. The study is made to constitute a system framework of mould coordinated manufacturing system formed by a distributed mould manufacture resources, mould cloud manufacturing service platform and the client-side based on the cloud manufacturing of cross-regional distribution mode design and manufacturing. It also explored the mould design manufacturing ideas and their implementation techonology under such new mould, and further describe the flow of work with the case analysis. The cloud manufacturing of cross-regional distribution mode design and manufacturing is of far researching importance on improving the availability of increasing the manufacturing resources and its design manufacturing capability, shortening the mould manufacturing period and reducing the costs of enterprises.

In order to solve the problem of BP neural network, genetic support vector regression is presented to predict the lifetime of cylinder. Support vector regression (SVR) is a novel prediction algorithm based on structure risk minimization principles, which can lead to great generalization ability. In the genetic support vector regression model, the genetic algorithm is used to optimize the parameters of support vector regression. That's because that the generalization ability of support vector regression is controlled by its parameters. The wear rate data of 20 mileages are employed to study the lifetime prediction of cylinder by the GSVR model. In order to prove the superiority of GSVR in lifetime prediction of cylinder, the RBF neural network and BP neural network are employed to compare with GSVR. The results of the experiments show that the lifetime prediction model of cylinder by the GSVR is better than that of RBF neural network, BP neural network.

In order to improve diagnostic accuracy and quality of maintenance, it is very important to study fault diagnosis method for automobile engine. Least-squares support vector machine called LSSVM is a modified SVM, which use a set of linear equations instead of a quadratic programming problem. In the paper, least-squares support vector machine is proposed to fault diagnosis of automobile engine. The LSSVM diagnostic model includes two LSSVMs which are used to recognize the three states of automobile engine including normal state, low-grade accidental fire and serious accidental fire. The experimental data of the relation between waste gas discharge and different accidental fire degree are presented to prove the diagnostic ability of the proposed method. The obtained results indicate that the used LSSVM method can make an effective interpretation in fault diagnosis of automobile engine.

Wideband spectrum estimation is very important for wideband wireless communication systems, especially for military communication system. The concept and the design method of the adaptive wideband spectrum estimation are put forward in this paper. Firstly, the wideband spectrum is estimated by using the wavelet packet decomposition. Then the characteristic noise of the independent signals exist in the wideband spectrum is produced by using the wavelet packet synthesis. Lastly the filters or trap filters, which are used to extract or trap signals, are designed based on LMS adaptive algorithm. The adaptive wideband spectrum estimation will be widely used in the wideband systems because of its real-time estimation ability and its good adaptability.

Vehicle flow detection plays an important role in ITS. In the process of vehicle flow detection, the vehicle is contiguous with another and the same vehicle counted repeatedly are common problems, especially the problem of changing lanes, which is very difficult to solve. This paper uses the method that combines background difference and virtual-loop sensor to detect vehicle flow, which based on the fast adaptive background updating method, and it introduces a new method of multi-lane sharing one detecting region. Experiment results show that this method can update the background exactly along with the variance of illumination; it also can solve vehicle orientation and vehicle changing lane problems. In real traffic scenes, it can detect vehicle flow accurately and meet the real-time processing demand.

Dynamic water surface simulation is an important part of simulating natural scene. Ben Humphrey used projection texture mapping, depth texture and mobile texture to render dynamic water surface and got a good result. In this paper, an improvement method will presented for speeding up its rendering speed and enhancing the rendering effects. It uses a viewpoint correlative method to replace depth texture for rendering the refractive effect and adds three parameters to control the direction of water movement, light intensity and size of waves. By interactive adjusting these parameters, the method can simulate the different states of the water surface. The experiment results show that this method can get faster rendering efficiency and more realistic effect.

Support Vector Machine (SVM), based on structural risk minimization principle, is now widely used in pattern recognition, classification and other research fields. It shows better generalization performance than traditional statistical learning theory, especially in small samples. In this paper, some dimensionless parameter is selected as SVM eigenvector, and then support vector machine is applied to fault diagnosis in engine parameter collector. Result shows that it has good ability in fault pattern classification of engine parameter collector.

Underwater optical effect simulation is an important part of simulating natural scene. However, computing these effects is difficult and time-consuming, since light refracts when passing through waves. To address the problem, this paper proposes a simple and quick method for rendering underwater scenes, including the shafts of light refracted by the dynamic water surface, caustics caused by the effects of light and shadows beneath the creatures in the water. The method simplifies the Iwasaki's method to save the computation-time by an eye-irrelative method calculating the intensities of the illumination volumes with parts of intersection points of Iwasaki's. As the experiment results show that our method got a high frame rate with an acceptable quality for real-time applications.

Cellular automaton and multiple scattering are the more important technologies for generating cloud in the phase of modeling and rendering. Cellular automaton model gets a balance between physical model and procedural model, which can make high speed and high quality clouds. Multiple scattering simulates the cloud scattering even more truly. In this paper, we will supply an algorithm which combines the above two technologies together to generate the real-time dynamic clouds. We hope using their strong suit to generate the real-time dynamic clouds. And through the cell's state-changing equation we model the cloud's movement by wind. Impostor is used to accelerate the rendering. All above get a better result of cloud's real-time movement and realization.