【作者】 张小龙；

【导师】 张为公；

【作者基本信息】 东南大学 ， 精密仪器及机械， 2006， 博士

【摘要】 论文研究针对装备底盘主动运动控制系统车辆行驶安全性能检测的整车道路试验系统及其性能评价方法。在分析了国内外有关汽车主动安全性法规及典型底盘主动运动控制系统关键技术的基础上提出了车轮力和车身位置姿态同步测量道路试验系统,并就测试方法、提高测量精度和性能评价方法等展开了系统深入的研究。论文主要研究工作概括如下:(1)基于车辆主动安全性研究的需要和国内汽车道路试验技术现状,提出了车轮力和车身位置姿态同步测量道路试验系统,研制了车身位置姿态测量模块,列出了各测量参数及其技术要求并对传感器选型。系统采用基于CAN总线的分布式数据采集和集中数据后处理的系统架构。位置姿态测量模块硬件设计选用了高集成、高速控制器,电源、信号和数传通道都采取了隔离措施;软件设计采用了模块化结构,对所有片上外设采用中断管理的方式。模块基于控制器自带32位定时器作为系统时基,较好地解决了卡内各通道信号、卡间数据时间同步问题。试验室调试和实车试验证明了该模块工作的可靠性。(2)在课题组车轮力传感器已有研究成果的基础上,从传感器标定方法和静态维间解耦角度深入研究了车轮力传感器维间耦合现象。对在自制液压标定试验台上标定的三分力数据分别采用多元线型回归、最小二乘支持向量回归机和改进正交最小二乘算法径向基神经网络三种方法进行解耦分析。结果表明:标定主通道线性检验显著;自研轮力传感器静态耦合率与国外产品精度相当;最小二乘支持向量回归机方法能够较好地克服小样本、非线性、高维数、局部极小点等实际问题且在回归精度、泛化能力和算法稳定性方面均具有良好的性能;对标准正交最小二乘径向基神经网络算法停止学习条件进行修正,并对扩展常数基于交叉验证法进行筛选,能够克服经验确定网络参数的不足,参数确定符合泛化基本理论,方法可行。(3)深入研究了中低精度捷联惯性测量系统和高精度载波相位模式差分GPS位置/速度/航向角间接反馈松散组合方式下车辆位置姿态测量的基本理论和工程实现关键问题。详细推导了汽车道路试验表述下的捷联惯性解算算法,基于惯性测量单元线性误差模型建立了21阶误差校正Kalman线性滤波器,并从该滤波器出发研究了现场快速动基座组合初始自对准方案。在kalman滤波器算法中,综合了平方根算法和自适应Sage-Husa算法,较好地克服了计算误差和模型不确定误差对滤波效果的影响。仿真和实车试验结果表明了该车辆位置姿态组合测量方法的可行性。(4)为了进一步完善测量模型和提高车轮力传感器的动态测量精度,综合考虑车辆主动安全性试验特点和模型实用性,建立了10自由度的整车模型,并基于牛顿力学原理进行了详细推导。该模型10自由度包括:整车纵向、横向线运动和横摆角运动,簧上质量的垂直线运动、侧倾和纵倾角运动,以及四个车轮的旋转运动,并考虑了空气阻力、轮胎滚动阻力、主动/半主动悬架(简化成等效弹簧和等效阻尼两部分)等因素。另外,基于该模型和车身位置姿态信息推导了联合工况下车身质心和车轮侧偏角、车轮滑动率等参数计算公式,是进一步对车辆主动安全性能评价分析的理论依据。探讨了模型中车身转动惯量和惯性积的实车路试测试方案。对比了基于车轮力和基于整车功率平衡两种测试轮胎滚动阻力系数方法,试验结果表明在滑行工况下两种方法表现出良好的一致性。(5)对ABS车辆制动稳定性能、评价指标、试验规程等进行了系统深入的研究。分析了目前制动法更多还原

【Abstract】 The active safety roadway test system is developed for the motor vehicle equipped with active movement control systems on the chassis in the paper. It is intended to evaluate the active safety in the light of the wheel forces, the path and attitude of the un-spring mass (PAUSM), and to establish the bridge between the driving performance and the vehicle structure, control parameters, etc. Based on both the analysis of the active safe laws and regulations, and the summarization of the key factors of these active movement control system, the roadway test system is put forth, whose primary features are to synchronize the measuring of the wheel forces and PAUSM. Several key points of the test system, such as measurement methods, measurement precision, performance evaluation, etc, are studied systematically. The primary work of the paper are summarized as follows:(1) It is for the wheel forces and PAUSM that the motor vehicle roadway synchronized measurement system is put forward primarily. The details of PAUSM module, the measurement parameters and requirements of the whole system are presented, and so do the determination of the sensor types . The system employs the structure of distributed parameters acquisition and centralized data post processing, and the CAN bus-based data management model between the main card and the slave card. Several means, such as isolation of signal channel, modular design of software, interrupt manage of the on-chip devices, etc., are utilized to improve the reliability of PAUSM module. Based on the 32-bit timer of the DSP chip, the time synchronization is easy to implement, both between main card and slave card, and among the different signal channel on PAUSM model. The reliability of PAUSM model is verified through the debugging on the lab and the test on the spot.(2) The coupling between each component of the wheel force transducer (WFT), is studied systematically in terms of both the calibration method and the static decoupling methods. Based on the calibrating data, three methods are utilized to analyze the coupling rate of WFT, which are the multi-component linear regression (MLR), the least square support vector regression (LS-SVR) and the improved orthogonal least square radial basis function neural network (OLS-RBF NN). Four points are generalized through the comparative analysis, (i) the linearity exists conspicuously on each main channel; (ii) the static coupling rate of the WFT is equivalent to the foreign like-kind products; (iii) the generalization and stabilization of LS-SVR are outstanding; (iv) the stop study condition of the normative OLS-RBF algorithm is modified, and the cross-valid method is employed to determine the net expansion constant. The improved algorithm, complied with the NN generalization theory and covering the shortage of empirical parameter determination, is verified effectively by the simulation analysis.(3) The principal theory and engineering technology of measuring PAUSM are approached respectively, based on the indirect feedback and non-tight integrated model between the middle-and-low-precision inertial measurement unit (IMU) and the high precision carrier-phase– differential GPS (CP-DGPS) receiver. Firstly, the strapdown algorithm is derived in detail under the frame of vehicle roadway test. Then the 21-order linear kalman filter is designed for the combination measurement of PAUSM based on the linear error model of IMU. Under the filter frame, the scheme of fast initial alignment on the spot is studied. The improved kalman evaluation algorithm,更多还原