Low-altitude UAV logistics networks in complex urban environments gain coordination through a collaborative two-stage optimization framework combining an improved Double Deep Q-Network with a risk-weighted Theta* algorithm to balance service loads and flight safety in Wuhan.
Urban delivery networks often struggle to coordinate transfer-node deployment with safe path planning. Dense populations and tall buildings complicate low-altitude airspace management. The new research addresses this coordination gap by merging macro-level node allocation with micro-level path generation.
Altitude-specific risk maps are constructed by incorporating population density, building height, and no-fly zones. These maps form the foundation for the upper-stage optimization process.
In the upper stage, a Double Deep Q-Network enhanced with NoisyNet and prioritized experience replay optimizes transfer-node selection and service allocation. The full model achieves a 22.3% boost in mean final reward and a 23.8% drop in mean episodes required to reach convergence relative to the standard DDQN.
The lower stage utilizes a risk-weighted Theta* algorithm to generate flight routes. This stage feeds path distances and risk costs back to the upper-level decisions.
Compared with conventional A*, risk-weighted Theta* reduces total service distance by 10.5%. It also cuts risk cost by 20.1% and total cost by 16.6%.
| Metric | Improvement |
|---|---|
| Mean Final Reward (vs. baseline DDQN) | +22.3% |
| Episodes to Convergence (vs. baseline DDQN) | -23.8% |
| Total Service Distance (vs. conventional A*) | -10.5% |
| Risk Cost (vs. conventional A*) | -20.1% |
| Total Cost (vs. conventional A*) | -16.6% |
Researchers conducted a case study in Wuhan to validate the framework. Results show that the developed architecture successfully harmonizes network configuration and risk-conscious path planning while simultaneously managing flight safety, service-load equity, and economic viability.
Wuhan, the capital of Hubei province with a population of 13,861,900, lies at the confluence of the Yangtze and Han rivers, a geography that historically earned it the name “Nine Provinces’ Thoroughfare.” Formed by the 1927 merger of Wuchang, Hankou, and Hanyang, the city has long functioned as a critical transportation hub where dense urban development and complex riverine topography create challenging low-altitude airspace conditions.
These results provide methodological support for UAV logistics network planning in complex urban environments.