A cost-effective mapping platform for digitizing the world: from autonomous vehicles to digital twins and the metaverse

数字孪生是指创建物理对象的虚拟副本，并在它们之间建立连接。数字孪生的概念已被嵌入到自动驾驶汽车、地理信息系统（GIS）、协作虚拟空间等现代创新的结构中，这些技术具有巨大的经济影响，预计到2026年，自动驾驶行业的价值将达到640亿美元。最近，Meta承诺在未来几年内投资100亿美元创建虚拟世界，预计将启动价值数千亿美元的创意经济。所有这些应用所需的骨干技术基础设施是真实世界/数字孪生模型的精确3D表示。大多数使用地面激光雷达扫描仪生成测量级地图的测绘技术都非常昂贵和耗时。相反，移动的地图绘制平台除了价格昂贵之外，还不能生成测量级地图。为了大规模地创建数字孪生模型，这对于自动驾驶和组合导航至关重要，地图平台需要低成本，同时还需要提供所需的准确性。我们提出了一个低成本的地图平台，包括相机，激光雷达，GNSS接收器，惯性传感器，可以安装在任何平台（汽车，背包，无人机）来绘制环境。在可能的资助期限内将开展的活动有：-系统校准：所有传感器的实验室和在线校准-定位：在GNSS和GNSS拒绝环境中的地理注册定位-从下游应用的原始数据中进行对象检测和分割-地图生成和3D重建：从注册数据中创建3D地图。该平台将能够在GNSS和GNSS拒绝环境中生成地理注册的测量级地图，这些地图也与自动驾驶车辆所需的高清地图兼容。

The term digital twin is to create a virtual copy of the physical object, with a connection between them. This concept of the digital twin has been embedded into the fabric of modern innovations such as autonomous vehicles, geographical information systems (GIS), collaborative virtual spaces, etc. These technologies have an enormous economic impact with the autonomous industry expected to be worth $64 billion by 2026. Recently Meta has promised to invest $10 billion into the creation of metaverse over the next few years which is expected to start a creative economy worth hundreds of billions of dollars. The backbone technological infrastructure needed for all these applications is a precise 3D representation of the real-world/digital twin.Most mapping technologies by using terrestrial LiDAR scanners for survey-grade map generation are extremely expensive and time-consuming. On the contrary, mobile mapping platforms cannot generate survey-grade maps in addition to being expensive. To create a digital twin at scale, which is essential for autonomous driving and combined workspaces, the mapping platform needs low cost while also providing needed accuracy. We propose a low-cost mapping platform consisting of cameras, LiDAR, GNSS receiver, an inertial sensor that can be mounted on any platform (automobiles, backpacks, UAVs) to map the environment. The activities that will be performed during the possible funding duration are, - System calibration: Laboratory and online calibration for all the sensors - Localization: Geo-registered localization in both GNSS and GNSS denied environments- Object detection and segmentation from the raw data for downstream applications - Map generation and 3D reconstruction: Creating a 3D map from the registered data. The platform will be able to produce geo registered, survey-grade maps that are also compatible with the HD Maps required by the autonomous vehicles, in both GNSS and GNSS denied environments.