智能车环境下车辆典型行为识别方法研究

发布时间：2018-01-04 06:32

本文关键词：智能车环境下车辆典型行为识别方法研究　出处：《长安大学》2015年硕士论文　论文类型：学位论文

【摘要】：智能车辆的行为识别作为智能车辆环境感知的重要组成部分,能够为智能车辆的控制和决策提供必要的数据支撑,是智能车辆安全、可靠运行的前提和基础。论文针对智能车辆典型行为识别问题,搭建了一种智能车辆典型行为识别系统,并在此基础上对智能车辆的典型行为识别方法进行研究。主要工作包括:(1)在对智能车辆体系结构和功能结构进行研究的基础上,构建了一种智能车辆典型行为识别系统,进而分析了该系统中各个模块的工作原理及其涉及的关键技术,并完成了系统的硬件选择。(2)在对智能车辆姿态估计算法进行深入研究的基础上,选择旋转矢量法作为陀螺仪的车辆姿态估计方法,选择高斯牛顿法作为加速度计和磁强计的车辆姿态估计方法,然后使用扩展卡尔曼滤波算法对两种估计方法的结果进行融合,实现了车辆姿态估计;通过仿真实验验证了该方法能够实现较好的车辆姿态估计效果。(3)针对车辆加速度积分求取车辆速度时的累积误差问题,提出了一种基于加速度修正的智能车辆速度估计方法。通过仿真实验验证了该方法能够有效减少车辆加速度误差和噪声对速度估计的影响,具有较好的车辆速度估计效果。(4)通过对智能车辆典型行为中车辆姿态和速度表现特点的分析,对车辆典型行为类型进行了划分;根据不同车辆行为类型的特点,分别提取车辆速度曲线的数字特征和车辆姿态信号的FFT系数作为车辆行为的特征,并将其作为支持向量机算法的输入量实现了智能车辆典型行为的识别。(5)搭建了一种智能车辆典型行为识别测试平台,完成了智能车辆典型行为识别软件和算法的设计,通过实验验证了所提出的智能车辆典型行为识别方法的有效性。
[Abstract]:As an important part of intelligent vehicle environment awareness, intelligent vehicle behavior recognition can provide necessary data support for intelligent vehicle control and decision-making, which is the security of intelligent vehicle. The premise and foundation of reliable operation. Aiming at the problem of intelligent vehicle typical behavior recognition, a typical behavior recognition system of intelligent vehicle is built in this paper. On this basis, the typical behavior recognition method of intelligent vehicle is studied. The main work includes: 1) on the basis of the research of intelligent vehicle architecture and function structure. In this paper, a typical behavior recognition system for intelligent vehicles is constructed, and then the working principle of each module and the key technologies involved in the system are analyzed. And completed the hardware selection of the system. 2) on the basis of the in-depth research on the intelligent vehicle attitude estimation algorithm, the rotation vector method is selected as the gyro vehicle attitude estimation method. Gao Si Newton method is chosen as the vehicle attitude estimation method of accelerometer and magnetometer, and the results of the two methods are fused by using extended Kalman filter algorithm to realize vehicle attitude estimation. The simulation results show that this method can achieve better vehicle attitude estimation effect. An intelligent vehicle speed estimation method based on acceleration correction is proposed. The simulation results show that the method can effectively reduce the impact of vehicle acceleration error and noise on speed estimation. By analyzing the characteristics of vehicle attitude and speed in the typical behavior of intelligent vehicles, the typical behavior types of vehicles are divided. According to the characteristics of different vehicle behavior types, the digital characteristics of vehicle velocity curve and the FFT coefficient of vehicle attitude signal are extracted as the characteristics of vehicle behavior. It is used as input of support vector machine algorithm to realize the recognition of typical behavior of intelligent vehicle. 5) A test platform for recognition of typical behavior of intelligent vehicle is built. The software and algorithm of intelligent vehicle canonical behavior recognition are designed, and the effectiveness of the proposed method is verified by experiments.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP391.41

【参考文献】