CRII: CSR: Empowering Sustainability and Intelligence into Air-ground Collaborative Systems

About This Grant

This NSF Computer and Information Science and Engineering Research Initiation Initiative (CRII) project will advance fundamental knowledge in intelligent and sustainable air-ground collaborative systems by integrating adaptive data compression, cooperative perception, and strategic planning. As drones and ground vehicles are increasingly deployed in critical applications such as search and rescue, environmental monitoring, and infrastructure inspection, there is a growing demand for autonomous systems that can operate effectively in dynamic, resource-constrained environments. While drones offer expansive visual coverage, they face limitations such as short operational endurance, constrained onboard processing capabilities, and high communication overhead when transmitting large volumes of video data. In contrast, ground vehicles provide greater computational power and more stable energy resources but are limited by narrow sensing ranges. This project addresses these limitations by developing a unified framework that enables seamless collaboration between aerial and ground platforms through efficient data handling, enhanced scene understanding, and adaptive mission planning. The expected outcomes will contribute to the development of energy-aware, resilient, and scalable autonomous systems that support public safety, environmental resilience, and national infrastructure. Educational impacts include integrating research outcomes into teaching and hands-on learning activities to strengthen student training in science, technology, engineering, and mathematics. The technical scope of the project involves three interconnected components. First, the project will develop a spatiotemporal compression framework using convolutional and reinforcement learning to dynamically adjust aerial video data according to scene complexity, thereby reducing bandwidth and energy demands. Second, it will create a lightweight transformer-based perception system to fuse information from aerial and ground perspectives, and apply a diffusion model to predict future environmental conditions. Third, the data collected from adaptive compression and environmental perception will be integrated into an optimization module that plans drone trajectories and charging schedules based on energy harvesting constraints. This coordination will enable strategic mission planning under uncertainty. The proposed framework will foster new insights into the development of intelligent and sustainable robotic systems for real-world deployment. 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.