There are an estimated 500 volcanoes that emit volcanic gases to the atmosphere. This project will design, build, and field-test a collaborative swarm of flying robots called the Volcano Co-robot Adaptive Natural algorithms (VolCAN) swarm. The VolCAN swarm will transform our ability to forecast volcanic eruptions. The swarm consists of multiple autonomous aerial drones that use algorithms inspired by biology to monitor the unpredictable environments surrounding volcanoes. In addition to monitoring gasses that precede volcanic eruptions, thereby protecting human lives, it will also measure how much carbon dioxide is emitted from volcanoes to better understand how they contribute to the global carbon budget. The VolCAN swarm can adapt to environmental conditions autonomously in real time, and it can also be guided by scientists to collect scientific data during the battery-limited flights of small drones. Our approach leverages the advantages of bio-inspired algorithms that are fast rather than perfectly accurate, and resilient rather than centrally controlled. The project will broaden participation in computing by involving students from underrepresented groups in both robotics research and programming courses.
This project will develop, analyze and rigorously test a co-robot swarm of unpiloted air vehicles (UAVs) that collect valuable scientific data in dynamic and unpredictable environments. The VolCAN swarm will use bio-inspired algorithms to detect CO2 plumes, descend plume gradients to measure maximum flux of CO2 from ground sources, estimate plume size, and infer maps of multiple CO2 sources over hundreds of square kilometers. Given battery limitations on flight times and dangerous, unpredictable conditions, the algorithms prioritize speed, robustness and interpretability over high accuracy. The novel bio-inspired algorithms scale to cover vast areas, adapt to the sensed environment to focus monitoring on the most important regions in real time, and are fast enough to collect many simultaneous emissions within the limited battery life of small UAV. Theoretical analyses will determine bounds on the speed and convergence times of the algorithms, and simulations and frequent field tests will measure the performance of the VolCAN system in rigorous, replicated experiments. Additionally, the project will demonstrate that scientists can operate the VolCAN swarm to collect data in the field with either fully autonomous adaptive surveillance or with scientist guidance. This approach combines human flexibility and judgment with the speed and mobility of a UAV swarm. The project will demonstrate the broad applicability of the VolCAN swarm in environmental monitoring applications in experiments to measure methane emissions from pipelines and assess ecological health of plant communities. It will also show that bio-inspired robots can function outside of highly-structured factories, labs, and warehouses to gather valuable scientific data in the hazardous and unpredictable environments of active volcanoes.
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.
