低速环境下的智能车无人驾驶技术研究

发布时间：2018-03-14 14:26

本文选题：智能车　切入点：无人驾驶　出处：《浙江大学》2015年硕士论文　论文类型：学位论文

【摘要】：智能车是一种涉及环境感知、人工智能、自动控制、车辆工程等多种学科的全新汽车概念和产品。无人驾驶作为智能车的关键技术和研究方向,已经引起了国内外诸多部门、企业、学校的高度重视,并取得了一定的研究成果。城市环境作为智能车的重点应用领域,由于其复杂的道路环境,成为了无人驾驶技术研究的重点和难点。本文以低速环境,特别是半结构化的城市交通环境作为研究对象,以实现无人驾驶相关关键技术作为研究目标。本文研究内容主要概括为以下几个方面：(1)对智能车和无人驾驶技术进行了深入调研,详细设计并分析了本文所用智能车的平台架构和无人驾驶方案,为本文无人驾驶关键技术的实现奠定了基础。(2)针对道路识别,本文提出了一种基于摄像头和激光雷达的道路边缘检测方法。该方法通过对图像处理结果和雷达分析结果进行数据融合可获得准确的道路边缘。(3)针对障碍物检测,本文提出了一种基于激光雷达的障碍物检测与避撞方法,该方法通过分析雷达数据获得障碍物所在的模糊区域,并对障碍物分布进行分析后得出最优行驶路线。(4)针对自主导航,本文提出了一种基于多传感器融合的地图导航方法。该方法通过对电子地图进行实时校正以获得正确的道路信息,利用GPS/DR组合定位获得精确稳定的位置信息,再结合雷达数据做出路口环境下的局部导航策略。(5)通过虚拟环境和真实环境下的无人驾驶实验,验证了本文所述无人驾驶技术的有效性和可靠性,并提出了本文所存在的不足之处和改进方向。
[Abstract]:Intelligent vehicle is a new vehicle concept and product involving environmental perception, artificial intelligence, automatic control, vehicle engineering and so on. As the key technology and research direction of intelligent vehicle, unmanned driving has caused many departments at home and abroad. Enterprises and schools attach great importance to it, and have achieved some research results. As a key application field of smart cars, urban environment has become the focus and difficulty of driverless technology research because of its complex road environment. Especially the semi-structured urban traffic environment as the research object, In order to realize the key technology of driverless, the research content of this paper is summarized as follows: 1) the intelligent vehicle and driverless technology are investigated deeply. The platform architecture and driverless scheme of the intelligent vehicle used in this paper are designed and analyzed in detail, which lays a foundation for the realization of the key technology of unmanned driving in this paper. In this paper, a road edge detection method based on camera and lidar is proposed. This method can get accurate road edge detection by data fusion of image processing results and radar analysis results. In this paper, a method of obstacle detection and collision avoidance based on lidar is proposed. By analyzing radar data to get the fuzzy area of obstacle, and after analyzing the distribution of obstacle, we can get the optimal route. In this paper, a map navigation method based on multi-sensor fusion is proposed, which can get correct road information by real time correction of electronic map, and obtain accurate and stable position information by GPS/DR combined positioning. Combined with radar data to make the local navigation strategy. 5) through the virtual environment and the real environment of unmanned driving experiments to verify the effectiveness and reliability of the unmanned driving technology described in this paper. The shortcomings of this paper and the direction of improvement are also put forward.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U463.6;U495

【参考文献】