【作者】 王晓霞；

【导师】 崔纳新；

【作者基本信息】 山东大学 ， 控制科学与工程， 2015， 硕士

【摘要】 汽车尾气是空气污染和温室气体的主要元凶,全球能源短缺也迫切需要可替代燃料用于汽车工业。在这种挑战下,汽车工业正在寻求一种能降低对传统能源的依赖,并改善汽车效率的替代驱动技术。在诸多解决方案中,Plug-In混合动力汽车(Plug-In Hybrid Electric Vehicle,简称PHEV)以其节能、环保及堪比传统汽车续驶里程的优势受到各汽车厂商的重视和用户的欢迎。为了充分发挥车辆动力总成系统的潜力,车辆能量管理系统的作用显得尤为重要。针对现有HEV能量管理策略只对单一性能指标进行优化,没有全面考虑其他性能指标,全局优化控制方法不能实时实现等局限性问题,本文提出考虑降低排放的粒子群多目标优化能量管理策略,不依赖于路况信息的限制,实现能量的合理分配。与基于确定性规则的能量管理策略相比,本文提出的策略在燃油经济性、排放性上表现均更加优异。无级变速器在性能以及结构上的优势使之成为混合动力汽车理想的传动装置。本文旨在解决汽车传动系统的冲击会对汽车零部件造成损害并且影响乘车体验这一问题,围绕无级变速器系统离合器计及驾驶性和NVH相关控制问题进行研究。总结影响离合器接合柔顺性的关键因素,设定发动机转速与负载信号作为输入变量,设计了离合器的模糊控制策略,提高起步稳定性,减小换档冲击度,优化转矩协调,实现乘车体验舒适度。对比试验的结果显示改善后结果均满足德国有关冲击度标准,仿真结果均体现了本文中离合器的模糊控制算法对车辆冲击度的有效改进。提出了一种基于驾驶者历史操作喜好的车辆主动安全系统。建立了计及驾驶动作特征和马尔科夫模型的驾驶行为预测方法。本文选取适当时长作为驾驶意图时窗,选取典型特性参数表征驾驶员行为与意图,通过Elman算法,建立适配的Elman模型并且利用实验数据对网络进行训练,借助于驾驶者的历史操作数据得到其在正常情况下的驾驶规律,从而预测驾驶员的下一步操作。仿真实验验证了训练得到的神经网络的有效性,并可以出色准确的反映驾驶者任意时刻、行车信息具有差异的操作规律。更多还原

【Abstract】 Vehicle industry is facing great challenge from energy shortage and also contributes to air pollution and comes with large quantities of global warming greenhouse gases. To reduce the reliance on the non-regeneration energy and improve running efficiency, Plug-In Hybrid Electric Vehicles (PHEV) are widely considered to be the most promising long-distance transportation vehicles instead of traditional engine vehicles.The energy management strategy has important theoretical significance on the performance of hybrid electric vehicle. However, the information of driving cycle is needed in advance in the global optimization, which is rarely practical. The energy management strategy based on the Particle Swarm Optimization (PSO) algorithm is designed to take both the vehicle’s fuel economy and minimum emission as objective function. Compare with deterministic Rule-Based energy management strategy, PSO algorithm can improve fuel economy and emission.CVT is an ideal transmission system for HEV, which has predominance at property and structure. The impulsion affects the endurance life of detail of vehicle and passenger ride comfort characteristic. The purpose of this paper is to solve above problems by is researching on the relevant problems concerning CVT vehicle clutch control and the improvements of drivability and NVH. Based on the design of the particle swarm algorithm and the proportional controller, key influence factors of the joint flexibility of the automatic clutch control are summarized. The engine speed and the load signal are used as the inputs, fuzzy control table is set up, relationships among the systematic parameters are derived, the binding extent of the clutch is determined, torque coordination control is realized and the ride performance is raised. The results shows degree of jerk can be reduced effectively and can be satisfactory to the requirement of German related regulations. Simulation results demonstrate the effectiveness of the proposed methods.The vehicle passive safety technology can only solve the problems caused by traffic accidents. The active safety technology, which can prevent and reduce accidents, would suffice for more far-reaching applications. In this paper Elman neural network is adopted to predict driver’s behavior ahead of time. The "people oriented" driver-vehicle-road closed loop model is set up. The system would record the habits of the driver and warn in time when the behaviors of the driver deviate from the forecasted trajectory to a certain extent. Real time simulation is carried out, which is based on 3D urban road that acquired by GPS equipment. The results indicate that Elman algorithm can be used to establish the warning system of driver’s improper operation and provide the reliable and valuable information for safe driving.更多还原