中文摘要：

采用虚拟实时控制器实现基于加速度信号的汽车悬架位移的实时测量.硬件测试系统采集器选用了工业级嵌入式系统CompactRIO及其C模块,传感器选用了ICP加速度传感器、拉线位移传感器和GPS系统.基于LabVIEW完成了FPGA编程、控制器编程,编程时采用了多线程和FIFO技术、采样数据打包时均包含了时间信息,有效提高了数据传输的快速性和数据处理的准确性.在控制器中实现了基于LabVIEW的小波滤波和支持向量机（SVM）模型的悬架位移实时预测输出（CAN,100 Hz）.根据悬架下测点的加速度均方根值对7种典型路面等级进行了划分,并将路面等级信息作为SVM模型的输入元素.对7种路面系统试验和数据分析表明：选用的ε-SVR算法,取ε=0.01时,控制器小波滤波一个点和预测一次悬架位移耗时小于1 ms,满足实时性要求;悬架位移模型预测曲线与实际测量曲线相关系数基本在0.90以上,满足精度要求.

英文摘要：

Real-time measurement for automotive suspension distance based on acceleration was carried out by using virtual real-time controller. This paper set up a test system, in which the industrial embedded system CompactRIO and several C modules were employed as its hardware collector, besides typical ICP acceleration sensor, displacement sensor and GPS system were also employed. FPGA programming and controller programming using the multi-threading and FIFO method were completed in LabVIEW. The time information was contained in every sample data package, which effectively improved the speed of data transmission and the accuracy of data processing. Wavelet filter and support vector machine （SVM） model were built and embedded in CompactRIO to predict and output suspension distance （CAN, 100 Hz） in real time. The grade of seven kinds of typical roads was determined according to the RMS value of acceleration of suspension＇ s lower measuring point, which was employed as input element of the SVM model. The road way tests and data analyses indicated that the controller consumes less than 1 ms when completing one point＇s wavelet filtering and predicting in which ε =0. 01 for ε-SVR algorithm. The relationship coefficient of the suspension displacement curves of predicted by SVR model and measured in road way tests were more than 0.90 primarily. So the system could meet the requirements of accuracy and real-time control.