基于GA-BP神经网络的汽车空气阻力系数预测研究

针对整车空气动力性能开发中数值计算耗时长的问题，提出一种基于GA-BP神经网络的汽车空气阻力系数预测方法。将汽车部分特征参数作为输入变量，经外流场仿真得到的空气阻力系数作为输出变量，获取数据集。采用遗传算法对BP神经网络进行参数寻优，最终建立基于遗传算法的BP神经网络模型，验证不同训练集数量对模型预测精度的影响。结果表明，GA-BP神经网络在训练样本较少时也能维持较高的预测精度，可用于汽车空气阻力系数的快速测。     

ISG柴油混合动力客车能量分配策略研究

以ISG柴油混合动力客车整车控制策略为基础,综合考虑客车燃油经济性和排放,提出一种基于模糊神经网络的能量分配策略,该策略根据模糊控制原理,结合人工神经网络自主学习功能,以电池SOC、整车需求转矩以及发动机转速为模糊输入来确定发动机和电机的最佳输出转矩分配,再以神经网络对控制的模糊规则进行记忆.仿真结果表明,该策略比普通逻辑控制更加有利于燃油经济性的提高,并在一定程度上改善了排放性能.     

基于神经网络的车用汽油机过渡工况空燃比辨识

以HL495Q电喷汽油机为研究对象,提出了一种基于BP神经网络的空燃比辨识方法,比较了不同拓扑结构的神经网络对空燃比辨识精度的影响,得到了一种最优的空燃比模型。试验结果表明,空燃比模型能高精度地逼近空燃比的实际动态过程,模型的平均相对误差小于2%。     

干道信号协调控制优化方法

针对干道信号协调控制中控制策略及其控制优化参数设置上存在的问题,考虑到智能控制中Sugeno模糊推理所具有的复杂系统动态性能表达能力和神经网络控制方法所具有的学习能力,结合两者的优势建立了1种基于模糊神经网络的交通信号协调控制模型,实现对绿信比与相位差的优化,使用微观交通仿真软件Vissim对干线交通进行了仿真研究,仿真结果表明,该方法能更为有效地减小平均车辆延误。     

汽车ABS中的模糊神经网络模型参考自适应控制策略

针对汽车制动的特点以及汽车防抱死制动系统的性能要求,建立了汽车的数学模型,提出一种模糊神经网络的自适应控制方案,构建了基于模糊神经网络的控制器和辨识器的结构模型。通过对网络参数的离线训练得出其初值,在控制过程中对网络参数进行在线微调,实现对汽车制动过程的有效控制。仿真结果表明:在不同的路面,汽车均能保持在最佳滑移率附近进行制动,制动时间及距离比较理想,满足ABS的安全性能要求。     

基于扰动观测的客车ESC自适应神经网络滑模控制

为了提高客车电子稳定性控制系统(ESC)的控制精度,针对实际车辆系统建模中存在各种非线性扰动项以及传统滑模控制(Sliding Mode Control,SMC)中抖振较大的问题,提出一种自适应神经网络滑模控制算法.基于2自由度车辆模型,首先设计一个二阶滑模(Second-order Sliding Mode,SOSM...     

基于GT—DRIVE的整车动力性燃油经济性仿真

汽车的动力性和燃油经济性是其最基本与最重要的性能之一。为在汽车开发之初分析整车的动力性和燃油经济性,文章研究了动力性和燃油经济性的评价指标,合理选取整车参数,利用GTDRIVE软件对整车进行动力学和运动学建模与仿真计算。将仿真结果与试验结果对比,合理设置边界条件修正模型,最终得到较为精准的模型,使得在没有样车阶段,利用GT—DRIVE软件仿真可以得到整车的动力性和燃油经济性参数,能够对动力系统的匹配方案进行定量评价,为以后汽车的动力系统匹配优化提供一定的理论参考。     

基于多目标遗传算法的混合电动汽车参数优化

动力系统和控制器参数的同时优化是提高混合电动汽车(HEV)燃油经济性并降低排放的关键。这类优化问题涉及多个相互冲突的优化目标和非线性约束,是典型的多目标优化问题。文中采用多目标遗传算法求解该优化问题的Pareto最优解集,并应用ADVISOR对实际算例的优化结果进行比较分析。结果表明,应用该方法可找到多组可行解,在满足原车动力性要求的前提下能有效提高燃油经济性,降低排放。     

基于粒子群算法的Plug—in混合动力汽车能量管理策略优化研究

基于对混合动力汽车能量管理策略优化的目的,建立了丰田PnusPlug-in混合动力汽车的MATLAB／Simulink数学模型,用数学公式描述了系统优化控制问题,采用粒子群优化算法对该包含众多约束条件的非线性优化问题进行了求解,利用PSAT专业软件对比分析了基本型优化控制算法、改进型优化控制算法和规则控制算法等的控制效果及燃油经济性。结果表明,经过优化后的Plug-in混合动力汽车在不牺牲汽车各项性能的前提下能提高动力系统工作效率。     

Simulation of a powertrain system for the diesel hybrid electric bus

The plug-in hybrid electric bus (HEB) is designed to overcome the vulnerable driving range and performance limitations of a purely electric vehicle (EV) and have an improved fuel economy and lower exhaust emissions than those of a conventional bus and convention HEBs. The control strategy of the plug-in parallel HEB??s complicated connected propulsion system is one of the most significant factors for achieving a higher fuel economy and lower exhaust emissions than those of the HEV. The proposed powertrain control strategy has flexibility in adapting to the battery??s state of charge (SOC), exhaust emissions, classified driving patterns, driving conditions, and engine temperature. Simulation is required to model hybrid powertrain systems and test and develop powertrain control strategies for the plug-in parallel HEB. This paper describes the simulation analysis tools, powertrain components?? models and modifications, simulation procedure, and simulation results.     

Mean Value WGT Diesel Engine Calibration Model for Effective Simulation Research

Recently, to improve vehicle fuel economy, as well as the performance of internal combustion engines, optimized system matching between a vehicle’s drivetrain and engine has become a very important technical issue. For this reason, the need for simulation research on engine and vehicle performance improvement has increased. But in general, since both engine simulation and vehicle simulation require initial engine calibration map input, a simple engine calibration method is required for the efficient configuration of various virtual engine calibration map setups. On this background, in this study, an example of waste gate turbocharger (WGT) cooled — exhaust gas recirculation (EGR) Diesel engine calibration using a test-based mean value engine model is presented as a suitable engine calibration map setting method. Also, the feasibility of an engine calibration model is confirmed through various engine tests. Using the simple model presented here, it is possible for diverse engine operating conditions and engine performance maps to be acquired.     

Economic value and utility of a powertrain system for a plug-in parallel diesel hybrid electric bus

This research is the first to develop a design for a powertain system of a plug-in parallel diesel hybrid electric bus equipped with a continuously variable transmission (CVT) and presents a new design paradigm of the plug-in hybrid electric bus (HEB). The criteria and method for selecting and sizing powertrain components equipped in the plug-in HEB are presented. The plug-in HEB is designed to overcome the vulnerable limitations of driving range and performance of a purely electric vehicle (EV) and to improve fuel economy and exhaust emissions of conventional bus and conventional HEBs. The control strategy of the complicated connected propulsion system in the plug-in parallel HEB is one of the most significant factors in achieving higher fuel economy and lower exhaust emissions of the HEV. In this research, a new optimal control strategy concept is proposed against existing rule-based control strategies. The optimal powertrain control strategy is obtained through two steps of optimizations: tradeoff optimization for emission control and energy flow optimization based on the instantaneous optimization technique. The proposed powertrain control strategy has the flexibility to adapt to battery SOC, exhaust emission amount, classified driving pattern, driving condition, and engine temperature. The objective of the optimal control strategy is to optimize the fuel consumption, electricity use, and exhaust emissions proper to the performance targets. The proposed control strategy was simulated to prove its validity by using analysis simulation tool ADVISOR (advanced vehicle simulator).     

Powertrain system optimization for a heavy-duty hybrid electric bus

This research concerns the design of a powertrain system for a plug-in parallel diesel hybrid electric bus equipped with a continuously variable transmission (CVT) and presents a new design paradigm for the plug-in hybrid electric bus (HEB). The criteria and method for selecting and sizing powertrain components equipped in the plug-in HEB are presented. The plug-in HEB is designed to overcome the vulnerable limitations of driving range and performance of a purely electric vehicle (EV), and it is also designed to improve the fuel economy and exhaust emissions of conventional buses and conventional HEBs. Optimization of the control strategy for the complicated and interconnected propulsion system in the plug-in parallel HEB is one of the most significant factors for achieving higher fuel economy and lower exhaust emissions in the hybrid electric vehicle (HEV). In this research, the proposed control strategy was simulated to prove its validity using the ADVISOR (advanced vehicle simulator) analysis simulation tool.     

基于ADAS联网时空数据的路段交通参数估算模型

交通参数实时获取是道路交通管控的重要基础.针对固定检测器观测范围受限和浮动车数量需求大的问题,研究了1种利用车载ADAS联网数据进行路段交通参数估算的方法.通过分析车载ADAS感知的前向目标参数与交通参数的关系,结合广义交通量定义,并考虑多车道条件下ADAS车辆及其邻近前车的相对运动变化特性,建立了1种非稳态交通条件下...     

基于预测恒定车头时距策略的自适应巡航控制研究

为了进一步提高车辆跟车过程中的跟踪性、安全性、舒适性和燃油经济性，针对已有间距策略表现过于保守或反应过于激烈等不足之处，提出了一种预测恒定车头时距策略。该策略考虑了相对加速度，建立了一种预测型期望车间距模型，进而应用于模型预测控制的多目标自适应巡航控制系统中，能进一步提高模型预测控制对多个控制目标的综合协调能力。搭建上层控制器、下层PID控制器、油门制动切换、逆纵向动力学模型。在多工况下仿真，通过建立性能评判指标对多目标进行量化分析。结果表明，所提出的间距策略在保证安全性的前提下，提升了自适应巡航控制系统的综合性能。在不同驾驶风格的车头时距下，跟踪性、舒适性和燃油经济性均有良好表现。     

一种提高轮边驱动客车经济性的驱动控制方法

轮边驱动电动客车采用4个永磁同步电机,通过减速器将驱动力传递至驱动轮。合适的转矩分配控制策略可以提升行车经济性。以轮边驱动电动客车为研究对象,采用加速踏板平滑处理和基于电机电动效率Map图的转矩优化分配方法,并通过AVL Cruise/Simulink联合仿真、dSPACE硬件在环和实车试验进行验证。结果表明,相比于平均转矩分配,采用加速踏板平滑处理和基于电机电动效率Map图的转矩最优分配方法可降低2.35%的能耗,且该控制算法在硬件在环和实车试验中有着较好的实时性,能够满足实车行驶的需求。     

灭火智能车辆PID控制方法研究

为满足灭火智能车辆在火灾现场所需的精确运动要求和技术性能,充分利用BP神经网络能够逼近任意非线性系统的优点,将BP神经网络和PID控制相结合,把训练后的网络输出作为PID控制器输入,并不断调整其P、I、D参数,进而调整控制器的输出电压以控制灭火智能车辆的速度。     

基于单神经元的汽车方向自适应PID控制

针对汽车方向动力学控制存在的非线性和参数时变不确定性问题，提出了一种新的基于单神经元的汽车方向自适应PID控制算法。该算法利用了神经网络的自学习和自适应能力，实现了方向PID控制器的参数在线自整定，从而避免了传统的自适应PID控制必须在线辨识被控系统的参考模型参数而带来的计算工作量大的问题。仿真计算和场地试验验证表明该控制算法可有效地控制汽车按照预期给定的轨迹行驶，且保证了汽车方向闭环控制系统具有较强的适应性和鲁棒性。     

增程式电动汽车动力系统参数匹配及控制策略优化

针对增程式电动汽车动力系统参数匹配的问题,在Simulink-Cruise联合仿真平台上建立了用于匹配设计的整车初始模型,提出了基于典型工况统计分析的匹配设计方法,对增程式动力系统进行了稳态匹配.为了进一步验证设计参数的合理性,采用恒温式定点控制策略和CD-CS型最优曲线功率跟随控制策略进行了仿真对比分析,验证了匹配参...     

基于REINFORCE算法和神经网络的无人驾驶车辆变道控制

针对无人驾驶车辆变道超车场景,研究基于REINFORCE算法和神经网络技术的无人驾驶车辆变道控制策略.通过车辆动力学模型确定模型的反馈量、控制量和输出限幅要求;设计神经网络控制器的结构,根据REINFORCE算法设计控制器训练方案;分析经验池数据数值和方差过大的问题,提出1种经验池数据预处理的方法以改进控制器训练方案;...