SBIR Phase I: Georeferenced Augmented Reality for Knowledge-Based Excavator Control

该项目具有更广泛的影响和商业潜力，因为它能够将挖掘机的操作和控制从主要以技能为基础的活动转变为以知识为基础的实践，从而显著提高生产率和安全性。这将有助于实现巨大的成本节约和减少对公众的潜在危害，提高美国建筑业的竞争力，降低民用基础设施的生命周期成本。在制造业、运输业、采矿业和造船业等领域，从以技能为基础的过程向以知识为基础的过程的转变是有价值的。从长远来看，拟议解决方案的意外公用事业罢工预警能力也有可能为客户节省数百万美元的责任和机会成本，避免对生活和商业的中断，并防止对工人和公众的人身危险。因此，预计该项目的社会效益和商业影响将是，通过安全和有效的挖掘，将可能大大减少建筑和地下基础设施的成本。这个小型企业创新研究（SBIR）第一阶段项目将把基于计算机视觉的基本运动跟踪技术转化为一种变革性的解决方案，用于在现场条件下跟踪挖掘机末端执行器（铲斗），并展示满足目标市场性能需求的能力。挖掘是一项典型的建筑活动，每个操作员都面临两个主要问题：1)需要保持精确的坡度控制；2)需要避免意外的公用事业罢工。该项目将克服这些痛点，并利用增强现实技术实时为作业者提供可视化的挖掘作业计划、目标品位剖面和不断变化的品位剖面，使作业者能够以高精度、更高的生产率和安全性实现目标品位。颠覆性创新是使用廉价的基于计算机视觉的跟踪：1)在局部坐标系中跟踪挖掘机铲斗的位置；2)在同一坐标系统中跟踪安装在舱内的摄像头的位置，以在增强现实中可视化隐藏的公用事业位置。大量的现场测试表明，该技术可以相对于相机跟踪标记物，不确定度小于1英寸，与目前基于全球定位系统的方法相比，具有显著的优势。

The broader impact/commercial potential of this project stems from its ability to transform excavator operation and control from a primarily skill-based activity to a knowledge-based practice, leading to significant increases in productivity and safety. This is turn will help realize enormous cost savings and reduction of potential hazards to the public, improve competitiveness of U.S. construction industry, and reduce life cycle costs of civil infrastructure. Such benefits will also accrue in fields such as manufacturing, transportation, mining, and ship-building where the transition from skill-based to knowledge-based processes is of value. In the long-run, the accidental utility strike warning capabilities of the proposed solution also have the potential to save millions of dollars in liability and opportunity costs for customers, and avoid disruptions to life and commerce, and prevent physical danger to workers and the public. The societal benefit and commercial impact of the project are thus expected to be the significant reductions in construction and underground infrastructure costs that will be possible through safe and efficient excavation. This Small Business Innovation Research (SBIR) Phase I project will translate fundamental computer-vision based motion-tracking technology into a transformative solution for tracking an excavator end-effector (bucket) in field conditions, and demonstrate the capability to meet target market performance demands. Excavation is a quintessential construction activity where every operator faces two major problems: 1) The need to maintain precise grade control; and 2) The need to avoid accidental utility strikes. This project will overcome these pain points and present operators with a visualization of excavation job plans, target grade profiles, and the evolving grade profile in real-time using augmented reality, allowing operators to achieve target grades with high precision, improved productivity, and safety. The disruptive innovation is the use of inexpensive computer-vision based tracking to: 1) track the position of an excavator bucket in a local coordinate system; and 2) track the position of a cabin-mounted camera in the same coordinate system to visualize buried utility locations in augmented reality. Extensive field testing has demonstrated that this technology tracks markers relative to a camera with an uncertainty of less than one inch, offering significant advantages over current global positioning system based methods that are expensive and unreliable for the pursued application.