CPS: Medium: Data-Driven Adaptive Real-Time (DART) Flow-Field Estimation Using Deployable UAVs

福岛第一核电站灾难和阿利索峡谷天然气泄漏是最近引人注目的由空气污染物意外释放导致的紧急情况的例子。在这种紧急情况下，对污染物运动的准确实时预测对于规划应急响应、保护应急工作人员和评估环境影响是非常宝贵的。然而，由于大气湍流、地面地形拓扑和不断变化的风力条件，对污染物扩散的准确预测是具有挑战性的。该项目通过使用自主无人驾驶飞行器（uav）车队获得大气流量和污染物浓度的稀疏物理测量，解决了实时预测大气污染物扩散的问题。然后，将这些稀疏的物理测量值实时用于不断改进计算流体动力学模型，以便对污染物的分散进行准确的实时预测。这代表了空气污染实时传感与多车辆群控制和云分散预测的紧密结合，以产生最佳车辆路径。该项目的主要目的是开发和演示一种新的数据驱动的自适应实时（DART）系统，该系统可以产生准确的实时微气象估计和预测污染源附近的污染物扩散。DART系统将包括一个计算流体动力学网络系统和一个由配备流量传感器和污染物浓度传感器的自主无人机组成的物理系统。总之，该DART系统将产生准确的流场估计，可用于预测污染物的分散。开发DART系统需要实时数据驱动模型适应的新技术，计算湍流建模的进步，基于无人机的传感和数据处理的改进，以及使用计算流体动力学网络系统的网络反馈的新型无人机编队飞行方法。该项目包括多个层次的实验，包括模拟、风洞和现场飞行演示，以提供概念验证。该项目由网络物理系统计划和促进竞争研究的既定计划（EPSCoR）共同资助。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The Fukushima Daiichi nuclear disaster and the Aliso Canyon natural gas leak are recent high-profile examples of emergency situations that resulted from the unplanned release of an airborne contaminant. In such emergency scenarios, accurate real-time prediction of contaminant movement is invaluable for planning emergency response, protecting emergency workers, and assessing environmental impact. However, accurate prediction of contaminant dispersion is challenging because of atmospheric turbulence, ground terrain topology, and changing wind conditions. This project addresses the problem of predicting atmospheric contaminant dispersion in real time by using a fleet of autonomous unmanned air vehicles (UAVs) to obtain sparse physical measurements of the atmospheric flow and contaminant concentrations. Then, these sparse physical measurements are used in real time to continually improve a computational fluid dynamic model in order to produce an accurate real-time prediction of the contaminant dispersion. This represents a tight integration of real-time sensing of airborne contamination with multi-vehicle swarm control and cloud dispersion prediction to generate optimal vehicle paths. The primary aim of this project is to develop and demonstrate a new data-driven adaptive real-time (DART) system that produces accurate real-time micro-meteorological estimates and forecasts contaminant dispersion near its source. The DART system will consist of a computational-fluid-dynamic cyber system and a physical system of autonomous UAVs instrumented with flow sensors and contaminant-concentration sensors. Together, this DART system will produce accurate flow-field estimates, which can be used to predict contaminant dispersion. Developing the DART system requires new techniques for real-time data-driven model adaption, advances in computational turbulence modeling, improvements in UAV-based sensing and data processing, and new UAV formation flying methods that use cyber-feedback from the computational-fluid-dynamic cyber system. The project includes multiple levels of experimentation including simulation, wind tunnel, and live flight demonstration to provide proof of concept. This project is jointly funded by the Cyber Physical System Program and the Established Program to Stimulate Competitive Research (EPSCoR).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.