AeroDaaS: A Programmable Drones-as-a-Service Platform for Intelligent Aerial Systems

A team of researchers from the Indian Institute of Science has unveiled AeroDaaS, a novel service-oriented framework designed to revolutionize how developers program and deploy intelligent drone systems. Published as an extended version of their IEEE ICWS 2025 paper, AeroDaaS addresses the significant challenge of orchestrating navigation, sensing, and analytics across drone, edge, and cloud resources. The platform introduces a Drone-as-a-Service (DaaS) model that abstracts the underlying complexities of UAV-based sensing, offering modular service primitives for on-demand sensing, navigation, and analytics as composable microservices. This approach ensures cross-platform compatibility and scalability across heterogeneous UAV and edge-cloud infrastructures, moving beyond the integrated, hard-to-manage designs that have hampered the field.

The technical implementation is remarkably efficient. AeroDaaS requires developers to write fewer than 40 lines of code to create applications, dramatically lowering the barrier to entry for aerial AI. The system introduces plug-and-play scheduling modules, including Waypoint and Analytics schedulers, which handle trajectory optimization and real-time coordination of inference workloads. In evaluations across six real-world DaaS applications (two in physical tests, four in simulation), the framework demonstrated minimal platform overhead: less than 20 milliseconds of latency per frame and approximately 1 gigabyte of memory usage on an NVIDIA Jetson Orin Nano and an AMD RTX 3090 GPU workstation. These results position AeroDaaS as a promising, flexible, and scalable foundation for the next generation of autonomous aerial analytics, enabling rapid development of applications from precision agriculture to disaster response.