A platform for high accuracy real-time tracking of dynamic multi-objects using auto-positioning and artificial intelligence to increase safety in construction sites

建筑业是加拿大所有行业中伤亡人数最高的行业之一。事实证明，实时定位系统（RTLS）能够通过提高工作人员的态势感知来缓解安全问题。超宽带（UWB）跟踪系统由于其高精度、高信号穿透特性和低功耗，在室内环境中显示出对动态物体进行精确跟踪的巨大潜力。然而，由于在安装时需要进行繁重的测量，需要电缆连接，并且成本高，因此在工作环境中广泛使用它们进行精确跟踪一直是一个长期存在的挑战。本研究的目标是开发一种准确、轻便、廉价且易于部署的平台，用于精确实时跟踪室内环境中的运动物体。该平台将使用廉价的、现成的双向测距UWB无线电，从而大大降低跟踪成本。双向测距消除了电缆连接的需要。自动定位算法用于消除安装时繁重的校准工作。一种新开发的滤波算法将用于提高自动定位的精度，使其达到与校准相似或更好的精度。机器学习和人工智能技术将用于克服时间延迟误差的影响，这在跟踪移动物体时很重要。所提出的位置估计平台的功能和精度将在实验室环境中进行测量，并将进行迭代校准以达到较高的精度。虽然该平台的直接最终用户是建筑行业，但在其他行业中，对动态资源（例如人员或设备）的实时态势感知可以提高安全性、安全性和生产力的行业中，对移动的多物体进行精确实时跟踪的潜力和需求很大；例如，跟踪医疗设施中的宝贵移动资源、跟踪老年设施中的患者或跟踪生产车间的设备。该平台可以为加拿大的各种行业提供竞争优势，包括建筑、制造业、医疗保健和机器人。

Construction has one of the highest rates of injuries and fatalities among all industries in Canada. Real-time locating systems (RTLS) have proven to be able to alleviate safety issues by increasing situational awareness of workers. Ultrawide Band (UWB) tracking systems have shown a high potential for precise tracking of dynamic objects in indoor environments due to their high accuracy, high signal penetration properties, and low power consumption. However, due to the requirement for onerous measurements at installation, the need for cable connectivity, and high cost, their widespread use for precision tracking in workplace environments has been a long-standing challenge. The objective of the proposed study is to develop an accurate, lithe, inexpensive, and easy-to-deploy platform for precise real time tracking of moving objects in indoor environments. The platform will use inexpensive, off-the-shelf two-way ranging UWB radios which drastically reduce the cost of tracking. The two-way ranging eliminates the need for cable connectivity. Auto-positioning algorithms are used to eliminate the need for onerous calibration at installation. A newly developed filtering algorithm will be used to enhance the accuracy of auto-positioning to the similar or better accuracy that is achieved with calibration. Machine learning and artificial intelligence techniques will be used to overcome the impact of time latency error that is important in tracking moving objects. The functionality and accuracy of the proposed location estimation platform will be measured in the lab environment, and it will be iteratively calibrated to achieve high accuracy. While the immediate end-user for the platform is the construction industry, there is a strong potential and demand for precise real-time tracking of moving multi-objects in other industries where real-time situational awareness of dynamic resources (e.g. people or equipment) can enhance safety, security and productivity; e.g. tracking valuable moving resources in healthcare facilities, tracking patients in senior citizen facilities, or tracking equipment on manufacturing floors. The platform can offer a competitive edge to various Canadian industries including construction, manufacturing, healthcare, and robotics.