一种基于BP神经网络的机动目标跟踪算法

交互式多模型（IMM）算法是一种可以有效跟踪机动目标的滤波算法,针对其跟踪精度和计算量在很大程度上受制于模型选择和转移概率确定的问题,提出了一种利用BP神经网络修正子模型滤波结果的改进IMM算法。仿真实验表明,该方法可以使IMM算法的收敛速度加快,收敛精度提高,改善了跟踪性能,具有一定理论指导意义。     

基于LVQ 神经网络的轨道单元状态综合评判方法研究

为了有效利用多种检测数据评判轨道单元的状态,提出利用 LVQ (学习矢量量化)神经网络建立轨道单元特征参数与轨道单元分级的关联模型,通过对 TQI (轨道质量指数)、轨道几何、加速度、晃车仪、添乘仪、人体感觉的超限扣分加权得到轨道单元的量化评分指标,并利用层次分析法确定各特征参数的权系数。根据大量实测数据建立随机样本,利用聚类方法确定轨道单元状态的分级。以轨道单元的量化评分指标作为输入,以聚类得到的表征轨道单元分级的矢量量化数据作为输出,利用误差反向传播方法训练 LVQ 神经网络模型。利用新的评判方法对某线路的轨道单元状态进行评判,结果表明该方法可行、有效,为轨道单元状态综合评判提供了一条新途径。     

一种基于语义分割结果的车道线检测拟合方法

车道线等地面标志物的检测是自动驾驶车辆环境感知的重要内容,能够为车辆提供可行驶区域的信息。文章提出一种基于语义分割结果的车道线检测拟合方法。使用车载单目相机获取车辆行驶过程中采集的道路图像,送入卷积神经网络进行车道线语义分割。将分割得到的仅含车道线的二值图像进行透视变换得到鸟瞰图,筛选有效车道线像素点,对有效车道线点使用最小二乘法进行多项式拟合,输出左右车道线多项式拟合系数,能够有效解决传统车道线检测算法的环境适应性差,鲁棒性不强,对弯道车道线检测信息不够准确等问题。     

Prioritizing road extension projects with interdependent benefits under time constraint

Since transportation projects are costly and resources are limited, prioritizing or sequencing the projects is imperative. This study was inspired by a client who asked: “I have tens of approved road extension projects, but my financial resources are limited. I cannot construct all the projects simultaneously, so can you help me prioritize my projects?” To address this question, the benefits and costs of all the possible scenarios must be known. However, the impacts (or benefit) of road extension projects are highly interdependent, and in sizable cases cannot be specified thoroughly. We demonstrate that the problem is analogous to the Traveling Salesman Problem (TSP). Dynamic change in travel demand during construction is another aspect of the complexity of the problem. The literature is yet to provide efficient methods for large cases. To this end, we developed a heuristic methodology in which the variation of travel demand during the construction period is considered. We introduce a geometrical objective function for which a solution-finding policy based on “gradient maximization” is developed. To address the projects’ interdependency, a special neural network (NN) model was devised. We developed a search engine hybridized of Ant Colony and Genetic Algorithm to seek a solution to the TSP-like problem on the NN based on gradient maximization. The algorithm was calibrated and applied to real data from the city of Winnipeg, Canada, as well as two cases based on Sioux-Falls. The results were reliable and identification of the optimum solution was achievable within acceptable computational time.     

DWL-48连续走行捣固稳定车控制系统分析

对DWL-48连续捣固稳定车的捣固装置控制系统进行分析,认为控制策略的滞后是导致设备现场使用故障的重要原因。以捣固架提升与下降控制为例,提出基于BP神经网络的捣固装置控制系统的设计思路。     

Kohonen神经网络的VC实现及测试

Kohonen神经网络能够模仿人脑特征进行自组织学习,并能根据其学习规则,对输入模式自动进行分类。文章通过对Kohonen神经网络模型的研究,运用VC语言实现其算法,并完成对该神经网络分类功能的测试。     

A Comparative Study of Three Multivariate Short-Term Freeway Traffic Flow Forecasting Methods With Missing Data

Short-term traffic flow forecasting is a critical function in advanced traffic management systems (ATMS) and advanced traveler information systems (ATIS). Accurate forecasting results are useful to indicate future traffic conditions and assist traffic managers in seeking solutions to congestion problems on urban freeways and surface streets. There is new research interest in short-term traffic flow forecasting due to recent developments in intelligent transportation systems (ITS) technologies. Previous research involves technologies in multiple areas, and a significant number of forecasting methods exist in the literature. However, most studies used univariate forecasting methods, and they have limited forecasting abilities when part of the data is missing or erroneous. While the historical average (HA) method is often applied to deal with this issue, the forecasting accuracy cannot be guaranteed. This article makes use of the spatial relationship of traffic flow at nearby locations and builds up two multivariate forecasting approaches: the vector autoregression (VAR) and the general regression neural network (GRNN) based forecasting models. Traffic data collected from U.S. Highway 290 in Houston, TX, were used to test the model performance. Comparison of performances of the three models (HA, VAR, and GRNN) in different missing ratios and forecasting time intervals indicates that the accuracy of the VAR model is more sensitive to the missing ratio, while on average the GRNN model gives more robust and accurate forecasting with missing data, particularly when the missing data ratio is high.     

船舶航向自适应神经网络鲁棒跟踪控制

针对存在不确定性和带有完全未知时变环境扰动的船舶航向非线性控制系统,将指令滤波技术和反步法相结合,设计了一种船舶航向自适应神经网络鲁棒跟踪控制器,该控制器的提出不依赖于外部扰动任何的先验信息。首先,利用神经网络补偿船舶航向非线性控制系统中的非线性项;设计自适应律在线更新神经网络权重向量和估计时变环境扰动的未知界。为了避免传统反步法中对虚拟控制律的反复求导,引入指令滤波技术,使得所设计的航向自适应神经网络跟踪控制器具有结构简单、易于工程实现的特点。理论分析表明,所设计的控制器能使船舶实际输出航向以任意期望的精度跟踪给定的参考航向,保证船舶航向闭环控制系统中所有信号一致最终有界。最后,以大连海事大学远洋实习船“育龙”轮为例进行仿真,仿真结果验证了所提控制器的有效性和鲁棒性。     

树脂混凝土强度预测的神经网络模型

本文运用均匀设计的方法进行树脂混凝土的配合比设计,用较少的试验取得较好的效果;建立了树脂混凝土的强度预测的神经网络模型,对试验的数据进行了训练和仿真,预测的结果与试验结果吻合非常好。     

混凝土结构钢筋锈蚀率BP神经网络预测模型

文章基于对钢筋锈蚀率与锈胀裂缝关系的分析,建立了根据锈胀裂缝宽度预测钢筋锈蚀率的BP神经网络模型,并将该模型的预测结果与采用线性拟合法预测的结果进行了比较。结果表明：用BP神经网络模型可以较准确、简便、快速地预测钢筋混凝土结构的钢筋锈蚀率。