SBIR Phase I: Intelligent Drone Ignitions To Manage Fires

该项目更广泛的影响/商业潜力是开发基于无人机的技术，以更安全，高效和有效地管理火灾。 野火每年在美国造成数千亿美元的损失。随着野火规模和数量的增加，规定的火灾已成为最具成本效益的工具之一，通过消除野火的多余燃料并控制其演变来减少灾难性损害。然而，为了最有效，规定的火灾需要将人放在火灾附近，并且通常受到无法获得（由于成本或需求）快速交付机制（例如直升机提供的快速交付机制）的限制。这一努力将通过提高无人机火灾监测和点火系统的智能来直接解决这两个痛点。该项目有可能塑造未来的火灾管理方式，使消防人员更安全，并实现具有成本效益的规定消防实践，可以帮助社会控制灾难性的野火，更好地管理景观。这个小企业创新研究（SBIR）第一阶段项目将使基于无人机的技术的开发能够改变火灾管理。目前基于无人机的点火和监控系统有可能改变这一领域，但由于广泛依赖人类控制器而受到阻碍。这种自动化的缺乏限制了这些系统？潜在的适用性，规模，有效性，并不必要地增加其风险。拟议的工作旨在通过以下方式使监测和点火更加智能和可扩展：1）专门的机器学习管道，以表征火灾前沿并识别地面资产; 2）轻量级火灾模拟引擎，可以在有限的计算机功率下执行时提供火灾前沿演变的短期预测;和3）分布式规划算法，可调整无人机的轨迹和球体投放策略，以操纵火力强度，同时保持地面资产、车辆和人员的安全。考虑到恶劣的火灾环境、商用无人机的重量和功率限制以及无人机导航和火灾管理两个不同领域的整合，这些挑战尤其难以应对。该奖项反映了NSF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 broader impact/commercial potential of this project is the development of drone-based technology to manage fires more safely, efficiently, and effectively. Wildfires cause hundreds of billions of dollars in annual damages in the United States. As the size and number of wildfires increases, prescribed fires have emerged as one of the most cost-effective tools to reduce catastrophic damage by removing excess fuel for wildfires and controlling their evolution. Yet, to be most effective, prescribed fires require putting people in the proximity of fire and are often limited by the unavailability (due to cost or demand) of fast delivery mechanisms such as those provided by helicopters. This effort will result in technology that directly addresses both pain points by improving the intelligence of drone-based fire monitoring and ignition systems. This project has the potential to shape how fire management is conducted in the future, making it safer for fire personnel and enabling cost-effective prescribed fire practices that can help society to control catastrophic wildfires and better manage landscapes.This Small Business Innovation Research (SBIR) Phase I project will enable the development of drone-based technology that will transform fire management. Current drone-based ignition and monitoring systems have the potential to change the field but are hampered by their extensive reliance on human controllers. This lack of automation limits these systems? potential applicability, scale, effectiveness, and unnecessarily increases their risk. The proposed work aims to make monitoring and ignition more intelligent and scalable through: 1) specialized machine learning pipelines to characterize fire fronts and identify ground assets; 2) light-weight fire simulation engine that can provide short-horizon prediction of the fire front evolution while executing on limited computer power; and 3) distributed planning algorithms that adjust the drone's trajectory and sphere drop strategy to manipulate fire intensity while keeping ground assets, vehicles, and personnel safe. These challenges are especially difficult to address given the harsh fire environment, the weight and power constraints of commercial drones preferred for these activities, and the integration of two distinct domains, drone navigation and fire management.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.