【摘要】 车辆定位导航系统实时、准确获取车辆位置对实现智能驾驶具有重要意义。针对传统定位导航系统存在的精度低、成本高、鲁棒性差等问题,基于GPS/INS(global positioning system/inertial navigation system,全球定位系统/惯性导航系统)、机器视觉和超声波雷达技术,设计了一种多传感器融合的智能车定位导航系统,旨在实现智能车在简单、结构化道路环境下的自动驾驶。利用GPS/INS技术实现智能车地理坐标获取,利用机器视觉技术实现智能车前方车道线检测,利用超声波雷达技术实现道路边沿检测,并对地理坐标、车道线和道路边沿数据进行深度融合,实现车道级定位导航。最后,进行了智能车定位导航现场测试,结果表明该系统满足车道级定位导航性能要求。研究结果表明,在简单、结构化道路环境下,多传感器融合的智能车定位导航系统结构简单,实际运行状况良好,可极大提高定位导航精度。更多还原

【Abstract】 In vehicle positioning and navigation system, real-time and accuracy of positioning are key indexes for intelligent driving. However, there are many problems in traditional positioning and navigation system, such as poor accuracy, high cost and poor robustness and so on. In order to solve these problems above, based on GPS/INS(global positioning system/inertial navigation system),machine vision and ultrasonic radar technology, an intelligent vehicle positioning and navigation system based on multi-sensor fusion(MSF) was designed, which could make intelligent vehicle automatic driving in simple and structured road. GPS/INS technology was used for obtaining geographic coordinates, machine vision technology was used for lane line detection and ultrasonic radar technology was used for road edge detection. What’s more, in order to achieve lane-level positioning and navigation, the data of geographic coordinate, lane line and road edge was fused by this system. Finally,the intelligent vehice positioning and navigation field test was carried out. The result showed that this system met the requirements of lane-level positioning and navigation. As to simple and structured road,this intelligent vehicle positioning and navigation system based on MSF with a simple structure runs well. Therefore, this design will greatly improve the accuracy of positioning and navigation.更多还原