Research on optimized GA-SVM vehicle speed prediction model based on driver-vehicle-road-traffic system

The accurate prediction of vehicle speed plays an important role in vehicle’s real-time energy management and online optimization control. However, the current forecast methods are mostly based on traffic conditions to predict the speed, while ignoring the impact of the driver-vehicle-road system on the actual speed profile. In this paper, the correlation of velocity and its effect factors under various driving conditions were firstly analyzed based on driver-vehicle-road-traffic data records for a more accurate prediction model. With the modeling time and prediction time considered separately, the effectiveness and accuracy of several typical artificial-intelligence speed prediction algorithms were analyzed. The results show that the combination of niche immunegenetic algorithm-support vector machine (NIGA-SVM) prediction algorithm on the city roads with genetic algorithm-support vector machine (GA-SVM) prediction algorithm on the suburb roads and on the freeway can sharply improve the accuracy and timeliness of vehicle speed forecasting. Afterwards, the optimized GA-SVM vehicle speed prediction model was established in accordance with the optimized GA-SVM prediction algorithm at different times. And the test results verified its validity and rationality of the prediction algorithm.     

Investigation on hierarchical control for driving stability and safety of intelligent HEV during car-following and lane-change process

The longitudinal and lateral coordinated control for autonomous vehicles is fundamental to achieve safe and comfortable driving performance. Aiming at this for hybrid electric vehicles (HEV) during the car-following (CF) and lane-change (LC) process while accelerating, a hierarchical control strategy for vehicle stability control is proposed. This new approach is different from the conventional hierarchical control. On the basis of model predictive control (MPC) theory, a two-layer MPC controller is designed at the top level of the control structure. The upper layer is a linear time-varying MPC (LTV-MPC), while the lower layer is a hybrid MPC (HMPC). For the LTV-MPC controller, a control-oriented linear discrete model for HEV is established, which integrates the dynamic model with three degrees of freedom (DOF) and the car-following model. The lower-layer HMPC controller is designed on the basis of the analysis for HEV hybrid characteristics and the modelling for the mixed logic dynamic (MLD) model of the HEV powertrain. As for the bottom level, a control plant including the HEV powertrain model and the 7 DOF nonlinear dynamics of the vehicle body is established. In addition, the system stability is proven. A deep fusion of vehicle dynamics control and energy management is achieved. Compared with LC-ACC control and conventional ACC control, the simulation and the hardware-in-the-loop (HIL) test results under different driving scenarios show that the proposed hierarchical control strategy can effectively maintain lateral stability and safety under severe driving conditions. Additionally, the HEV powertrain output torque and the gear-shift point are coordinated and controlled by the HMPC controller.

     

Modeling and optimal energy management of a power split hybrid electric vehicle

With the combination modes of engine and two electric machines,the power split device allows higher efficiency of the engine.The operation and of a power split HEV are analyzed,and the system dynamic model HEV is established event-driven for HEV forward system simulation dynamics controller design.Considering the mode,the fact the mode that the operation modes of is the are and the the is continuous theory.time-driven this for each structure selection of the controller built and the described finite with hybrid automaton control In control structure,process is depicted by the state mode machine(FSM).The multi-mode switch controller is designed to realize power distribution.Furthermore,vehicle operations programming are optimized,and finite the prediction nonlinear model horizon.predictive control(NMPC)strategy is applied by that implementing the dynamic(DP)and in the Comparative simulation The results optimal demonstrate strategy hybrid in control structure is effective feasible for HEV energy management design.NMPC is superior improving fuel economy.     

基于多头注意力卷积网络的电力负荷预测

预测居民用电相当于预测一个多元时间序列.针对多个传感器信号的特定窗口能够利用预测模型提取不同的特征来预测用电量.然而,由于时间序列内部特征存在着不规则的模式,包括电力属性之间隐藏的相关性,使得负荷预测准确率不高.为了提取复杂的不规则电力模式,选择性地学习时空特征以减少电力属性间的平移方差,本文提出了一种基于多头注意力的卷积循环神经网络深度学习模型.相较于单纯的时间序列模型,该模型利用卷积和加权机制对电力属性和有功功率间的局部相关性进行建模.它利用softmax函数和点积运算的注意力分数来模拟电力需求的瞬态和脉冲特性,有效地对瞬时脉冲功耗进行预测.在美国加州大学欧文分校(University of California, Irvine, UCI)家庭用电数据集共2 075 259个时间序列上的实验表明,所提出的模型与现有方法相比,准确率得到了较大提升.     

基于PCA - GRNN模型的新能源汽车月度销售量预测研究

为预测新能源汽车的月度销售量,提出了一种基于主成分分析(PCA)和广义回归神经网络(GRNN)相结合的预测模型——PCA - GRNN模型.首先,选取动力电池月份装车量、充电基础设施、电池级碳酸锂平均价格、交通和通信类居民消费价格指数、全国城镇调查失业率、汽车制造业工业生产者出厂价格指数等6个指标作为新能源汽车月度销售量的影响因子; 其次,利用主成分分析方法得到可代表6个影响因子的2个主成分,并利用Matlab神经网络工具箱的GRNN神经网络函数构建了广义回归神经网络模型; 最后,将2020—2022年间27个月度的统计数据分别输入到PCA - GRNN、PCA - BP和PCA - Elman模型中进行预测.结果显示, PCA - GRNN模型预测的新能源汽车月度销售量的平均相对误差(4.00%)低于PCA - BP模型和PCA - Elman模型预测的平均相对误差(分别为4.77%和4.29%),因此PCA - GRNN模型在预测新能源汽车销售量方面具有一定的实用性.     

An optimal energy management development for various configuration of plug-in and hybrid electric vehicle

Due to soaring fuel prices and environmental concerns, hybrid electric vehicle(HEV) technology attracts more attentions in last decade. Energy management system, configuration of HEV and traffic conditions are the main factors which affect HEV's fuel consumption, emission and performance. Therefore, optimal management of the energy components is a key element for the success of a HEV. An optimal energy management system is developed for HEV based on genetic algorithm. Then, different powertrain system component combinations effects are investigated in various driving cycles. HEV simulation results are compared for default rule-based, fuzzy and GA-fuzzy controllers by using ADVISOR. The results indicate the effectiveness of proposed optimal controller over real world driving cycles. Also, an optimal powertrain configuration to improve fuel consumption and emission efficiency is proposed for each driving condition. Finally, the effects of batteries in initial state of charge and hybridization factor are investigated on HEV performance to evaluate fuel consumption and emissions. Fuel consumption average reduction of about 14% is obtained for optimal configuration data in contrast to default configuration. Also results indicate that proposed controller has reduced emission of about 10% in various traffic conditions.     

洛阳市公共交通规划中的预测技术

文章介绍了“洛阳市公共交通规划”中所采用的预测方法和模型,着重论述了灰色动态预测 GM(1,1)模型的建模、模型处理及精度分析,并利用灰色关联分析和多元线性回归模型相结合进行预测,还提出了能开拓信息渠道反映预测技术灵活性的特征预测.     

应用灰色新陈代谢GM(1,1)模型预测河流水质

由于常规GM(1,1)模型进行预测时,精度较高的仅是最近的几个数据,越往未来发展,该模型预测的精度也就越弱。针对常规GM(1,1)模型存在的不足,运用灰色系统理论,建立了灰色新陈代谢GM(1,1)河流水质预测模型,对该模型的精度以及误差进行了分析,并利用该模型对某地区河流的水质进行了预测。计算机实际模拟证明:灰色新陈代谢GM(1,1)预测模型能够明显地提高预测精度,增加预测的可靠程度,从而实现河流水质的早期预测评估。     

基于Bi-LSTM的非等时距路基工后沉降滚动预测

为了实现路基工后沉降的早期、精准预测,提出基于双向长短期记忆网络（Bi-LSTM）的路基沉降预测技术. 采用Akima法将观测数据内插为适应时序分析法的等时距序列,提取“填土高度-时间-地基沉降”曲线中的6个影响因素作为变量训练Bi-LSTM模型,结合滚动迭代方法实现沉降预测的后延更新. 研究表明,利用深度学习技术可以有效地利用路基施工期信息,增加训练样本量,提升沉降早期预测的可靠性. Bi-LSTM模型对观测信息进行双向特征提取,同等样本量下的预测效果更精确. 依托6个中等压缩性土地基和1个复合地基监测断面信息,仅利用路堤填筑期及工后3个月数据,沉降预测的均方根误差 (RMSE) 和平均绝对百分误差 (MAPE) 平均值可以控制为1.19 mm、1.04%.     

系统效率最优的功率分流式混合动力汽车非线性预测控制

针对一种基于双行星排构型的功率分流式混合动力汽车,建立系统动态模型,准确描述其转速转矩耦合关系,通过建立各部件的效率模型,分析不同模式下系统的工作效率. 设计控制器结构框架,以系统工作效率和电池充放电平衡为目标,构建基于模型预测控制的优化问题,采用一步马尔科夫链模型预测驾驶员需求转矩及车速,将有限时域内的优化问题转化为非线性规划问题,基于序列二次规划算法实现优化求解. 仿真研究表明,基于系统效率最优的预测控制器能够维持电池的充放电平衡,在美国城市驾驶循环（UDDS）下,当电池初始电池荷电状态（SOC）分别为0.50、0.55和0.60时,相较于以发动机燃油消耗最优为目标,车辆等效燃油经济性分别提高了7.17%、5.73%和10.11%,验证了控制器的有效性和优越性.     

非平均权重条件下电力负荷组合预测模型研究

在对电力负荷进行预测过程中,单一的预测模型往往会出现拟合能力低和预测精度不高的缺陷,而且多个预测模型的简单组合也是一种相对粗略的预测方法。本文将非平均权重法运用于电力负荷组合预测模型中,以某省2000~2011年电力负荷为例,结果表明非平均权重的电力负荷组合预测模型的拟合方差比单项预测模型以及平均权重下的组合预测模型都小,而且预测精度更高。     

基于LSTM-KF的无人机航迹跟踪算法

在量测信息有限的情况下,针对使用单一运动模型的卡尔曼滤波(KF)算法难以应对无人机航道跟踪的问题,提出了一种新颖的将长短期记忆网络(LSTM)和KF算法结合的LSTM-KF算法。首先,使用LSTM预测目标平均速度和瞬时速度的方法解决了非参数模型在位置预测任务中泛化能力差的问题。其次,分析了KF算法使用运动模型的预测局限性,提出利用LSTM的预测结果修正运动模型的预测结果的方法,来降低预测误差。修正后的预测结果与量测数据结合,实现对目标的状态估计。最后,将所提LSTM-KF算法在生成的轨迹上进行了验证,仿真结果证明,LSTM-KF算法比已有模型具有更高的跟踪精度和更强的鲁棒性。     

认知NOMA系统中基于IRS的V2V网络物理层安全性能研究

认知非正交多址接入(NOMA, Non-Orthogonal Multiple Access)和智能反射面(Intelligent Reflecting Surface, IRS)由于其高频谱效率和低功耗而被认为是车联网两种有前景的技术。本文考虑恶意窃听者存在时基于认知NOMA的IRS辅助的车—车(Vehicle to Vehicle, V2V)网络,在不考虑信道估计误差的情况下,从安全性和可靠性两个角度研究了认知NOMA系统中基于IRS的V2V网络物理层安全性能,推导了双瑞利衰落信道下的中断概率和截获概率解析表达式,最后通过蒙特卡洛仿真进行了验证。结果表明,通过对源车辆发射功率、车辆间距离、IRS反射单元数量、目标速率以及功率分配系数等参数进行优化,能够进一步提升V2V网络的物理层安全性能。     

5G蜂窝网辅助的车载自组网数据传输机制与路由算法

针对车联网中不同种类数据的传输需求,该文提出一种V2V和5G蜂窝网络结合的混合消息传输机制及路由算法.将车联网中的数据包分为时延敏感型和非敏感型两种类型,利用5G蜂窝网低时延、高可靠性、网络覆盖范围广的优势,高效传输时延敏感型的数据消息.由于自组网比高性能的5G蜂窝网具有更低的成本,因此针对时延非敏感型数据包设计了一种...     

基于灰色RBF网络的我国能源消费预测

利用我国能源消费总量的历史数据,采用灰色预测法建立预测模型,再利用径向基（RBF）神经网络对灰色预测模型结果进行预测,以作为其最终的预测值．实验结果表明,灰色RBF网络模型在预测精度方面优于单一的灰色模型．该模型计算简便,有较高的拟合和预测精度,拓宽了灰色模型的应用范围．     

Co-simulation platform for single-shaft parallel HEVs based on Cruise+Matlab/Simulink

We reversely analyzed the energy management strategy （EMS） for a single-shaft parallel hybrid electric vehicle （HEV）, and build a forward co-simulation platform based on Cruise and Matlab. The vehicle dynamics model is built with Cruise, and control model is set up with Matlab/Simulink environment. The data between the two models are transferred by the Matlab API interface in Cruise. After mechanical and signal connections are completed, we establish the computing tasks and take the simulations of vehicle＇ s power performance, economy, and emission performance. The simulation results match the actual measurement results, which show that the co-simulation platform is correct and feasible. The platform can be used not only for a basic simulation platform to optimize further EMS, but also for the development of actual control system.     

深基坑工程监测数据处理与预测报警系统

介绍了“深基坑工程监测数据处理与预测报警系统”的功能、结构、开发运行环境及其建立方法 .该系统对深基坑的监测数据实施数据库管理 ;采用灰色系统理论建立变形预测模型 ;采用若干定性和定量指标进行深基坑工程极限状态的分析判别与险情预报 ;该系统是一个集监测数据处理、图形绘制与管理、变形预测及险情分析判别于一体的专家集成系统 .     

基于RS-ANN的某矿山井下开采防水安全岩柱厚度的确定

水下矿床是难采矿体的一个重要方面,针对某矿山的实际情况,深入分析了该矿的地质资料后,在查阅相关文献的基础上,建立了该矿导水裂隙带高度的粗糙集—神经网络预测模型.在比较了粗糙集—神经网络预测结果、地质详查报告提供的结果及采用经验公式计算的结果后,认为神经网络预测的结果较准确,其结论在该矿水下开采设计中可以采用.在预测的导水裂隙带高度基础上,参考采矿设计手册中的经验公式,并类比其它矿山水下开采的情况,计算出了防水安全岩柱的厚度.此项研究为该矿重新编制开采设计方案和安全专篇提供了依据,同时为矿山安全管理提供了参考.     

深基坑工程监测数据处理与预测报警系统

介绍了“深基坑工程监测数据处理与预测报警系统”的功能、结构、开发运行环境及其建立方法．该系统对深基坑的监测数据实施数据库管理;采用灰色系统理论建立变形预测模型;采用若干定性和定量指标进行深基坑工程极限状态的分析判别与险情预报;该系统是一个集监测数据处理、图形绘制与管理、变形预测及险情分析判别于一体的专家集成系统．     

组合模型在中期电力负荷预测中的应用

为了对中期电力负荷进行合理预测,结合三次指数平滑法和GM(1,1)预测模型,建立了新的组合模型,并以预测偏差平方和最小为准则优化了各单一模型的权重.通过MATLAB编程并以某市全年用电量为例对3种方法的预测精度进行了仿真验证.结果表明,组合模型具有更高的预测精度和更低的预测误差,能避免各单一预测模型的局限性.因此,用组合模型对未来用电量进行预测的结果更可靠.