Knowledge-aware path planning for UAV parcel delivery and road monitoring

Unmanned aerial vehicles (UAVs) are increasingly shaping the future of urban mobility, logistics, and infrastructure monitoring. Among their myriad applications, two domains have emerged as particularly impactful: parcel delivery and road monitoring. While delivery UAVs prioritize minimizing response times and optimizing routing for logistic efficiency, road monitoring UAVs focus on maximizing spatiotemporal coverage of transportation networks. This paper introduces a unified decision-making framework that enables UAV operators to simultaneously fulfill delivery requests and perform road monitoring in a coordinated manner. The urban environment is modeled as a graph-based road network with dynamic importance values assigned to road segments, which are refreshed upon a UAV visits and gradually increase over time in the absence of monitoring. To address this problem, we design a knowledge-aware rolling horizon planning framework that operates in a fully distributed manner, allowing each UAV to re-plan its trajectory iteratively based on local state and shared global knowledge. This distributed design makes the framework both scalable and easy to implement in practice. Candidate trajectories are generated by a lightweight search module, and the controller then selects and executes the most informative short-term action. Extensive simulations on real-world road networks demonstrate that our approach outperforms baseline methods in achieving a favorable trade-off between delivery performance and monitoring coverage.