水下智能潜器的神经网络运动控制

本文介绍一种基于神经网络的水下智能潜器的运动控制方法，该方法通过在线学习，融控制与滤波为一体。计算机仿真与水池实验验证表明，该方法的控制与滤波性能良好，对环境的学习与适应能力强。该方法事实上可用于一般动力系统的控制。     

一种基于BP算法学习的小波神经网络

为发展 Szu的基于信号表示的小波神经网络 ,提出一种多输入多输出的小波网络模型 ,网络隐层采用框架小波函数、输出层采用 Sigmoid激励函数 ,并选用“熵误差函数”以加速网络的学习速度。奇偶判别和混沌时间序列预测例子的实验结果表明了它具有良好的函数逼近能力和推广能力 ,收敛速度和均方误差均优于相同结构的多层感知器模型。     

Multi-scale variability of beach profiles at Duck: A wavelet analysis

Beach profiles have been observed to change over a range of spatial and temporal scales; however techniques for quantifying this variability have not been fully established. In this paper, a wavelet technique is introduced as a method to study the multi-scale variability of beach profiles. The beach profile data comprising a 22-year time series surveyed at the US Army Corps of Civil Engineers Field Research Facility (FRF) at Duck are analysed using the adapted maximal overlap discrete wavelet transform (AMODWT). The analysis successfully identifies strong local features in the variability of beach profiles in time and space separately that cannot be isolated by traditional statistical methods. The analysis of spatial wavelet variances provides a new means of investigating the depth of closure. Analysis of variances by temporal scales shows that the combined effects of several temporal scales with one or two dominant scales can be seen at particular points across profiles whilst the dominant temporal scales are different at different portions of the profiles. The method allows for the extremely nonstationary behaviour of beach profile to be analysed into separate frequency bands that can facilitate the interpretation of morphological changes in terms of physical processes.     

Inverse reconstructions of ecosystem flows in investigating regime shifts: impact of the choice of objective function

Inverse analysis is increasingly used in ecosystem modelling to objectively reconstruct a large number of unknown flows or interactions from a small number of observations. This type of analysis may be useful in relating observed regime shifts in ecosystem structure to underlying processes. Inversions of ecosystem flow networks currently use a constrained least-squares solution which at the same time minimizes the squared norm (the sum of squares) of the reconstructed flows. This minimum norm (MN) inversion is thought to be a parsimonious solution to the ecosystem flow inverse problem, but it may well not reflect how ecosystems are organised. It has been proposed instead that ecosystems evolve to maximize energy/mass flows or that they maximize the information content of the network weighted by ecosystem flows (ascendancy). We used simulated inverse experiments, where inverse analyses are applied to simulations of flow networks, to explore objective functions different than the MN generally used. We could not compute inverse solutions that maximize ascendancy because the objective function is unbounded. We could calculate inversions that maximize flows; however, these generally overestimated the simulated flows, even though the simulations were designed to maximize flows. It appears that the ecosystem flow inverse problem is too under-determined (too few data relative to the number of unknowns) to allow the use of these maximizing goal functions. We introduce a new minimization that simultaneously minimizes the squared flows and the squared differences between flows. This smoothing minimization makes the inverse flows as even as possible and it helps with some technical issues with MN inversions. The simulated inverse experiments indicated that this smoothed norm (SM) is the most robust in comparative analyses of contrasting ecosystem states, such as those that can be associated with regime shifts. Like the MN inversion, the SM inversion has no ecological basis. However, it is a conservative norm that is less likely to produce false differences between the dynamics of regimes.     

Response-based method for determining the extreme behaviour of floating offshore platforms

The accurate prediction of extreme excursion and mooring force of floating offshore structures due to multi-variete environmental conditions which requires the joint probability analysis of environmental conditions for the worst case situation is still impractical as the processing of large amount of met-ocean data is required. On the other hand, the simplified multiple design criteria (e.g. the N-year wave with associated winds and currents) recommended by API known as traditional method does lead neither to the N-year platform response nor to the N-year mooring force. Therefore, in order to reduce the level of conservatism as well as uncertainties involved in the traditional method the response-based method can be used as a reliable alternative approach. In this paper this method is described. In order to perform the calculations faster using large databases of sea states, Artificial Neural Networks (ANN) is designed and employed. In the paper the response-based method is applied to a 200,000 tdw FPSO and the results are discussed.     

葫芦岛地区道路网络的建立

本文针对城市交通道路网的特点,根据现实的交通网络,对葫芦岛地区道路网络进行建立及分析。利用ArcGIS中的Geodatabase构建路网数据库,利用节点一弧段数据结构构建空间网络拓扑关系,生成点一弧矢量数据模型表示的道路网络。并实现四种网络分析功能:最佳路径分析、空间查询、追踪分析和缓冲区分析。     

CPⅢ测量数据处理系统开发若干关键技术研究

本文首先介绍了无砟轨道的相关内容,指出CPⅢ测量数据处理是整个建设中的关键,提出系统开发的总体目标与要求且设计了系统总体结构,详细阐述了系统开发中涉及的关键技术与基本理论,文中最后结合具体工程实例验证系统开发成果。     

A data-driven approach to local gravity field modelling using spherical radial basis functions

We propose a methodology for local gravity field modelling from gravity data using spherical radial basis functions. The methodology comprises two steps: in step 1, gravity data (gravity anomalies and/or gravity disturbances) are used to estimate the disturbing potential using least-squares techniques. The latter is represented as a linear combination of spherical radial basis functions (SRBFs). A data-adaptive strategy is used to select the optimal number, location, and depths of the SRBFs using generalized cross validation. Variance component estimation is used to determine the optimal regularization parameter and to properly weight the different data sets. In the second step, the gravimetric height anomalies are combined with observed differences between global positioning system (GPS) ellipsoidal heights and normal heights. The data combination is written as the solution of a Cauchy boundary-value problem for the Laplace equation. This allows removal of the non-uniqueness of the problem of local gravity field modelling from terrestrial gravity data. At the same time, existing systematic distortions in the gravimetric and geometric height anomalies are also absorbed into the combination. The approach is used to compute a height reference surface for the Netherlands. The solution is compared with NLGEO2004, the official Dutch height reference surface, which has been computed using the same data but a Stokes-based approach with kernel modification and a geometric six-parameter “corrector surface” to fit the gravimetric solution to the GPS-levelling points. A direct comparison of both height reference surfaces shows an RMS difference of 0.6 cm; the maximum difference is 2.1 cm. A test at independent GPS-levelling control points, confirms that our solution is in no way inferior to NLGEO2004.     

变形分析的神经网络技术应用实例

大型工程施工过程中的变形监测、分析与预报极为重要。变形分析建模的方法很多,神经网络技术的应用是其中之一。文中结合某深基坑工程的监测资料和工作经验,运用神经网络BP算法进行预测分析。简述BP神经网络的基本概念,介绍基坑变形分析的BP神经网络的具体模型结构,将神经网络方法预报结果与实测数据对比效果较好。该成果对生产实践具有参考价值。