Real-time risk assessment of logistics drones considering uncertainty and multi-source risk factors

Multi-source risk factors in complex low-altitude airspace significantly affect the operational safety of logistics drones. Reliable quantitative risk assessment techniques are key to addressing this problem. This work proposes an improved risk field model based on field theory. The proposed model considers multi-source risk factors in the operation of drones from both the perspectives of dynamic drones and static buildings. A distance-sensing module and a speed and distance correction module were introduced into this model to achieve three-dimensional risk assessment. The proposed model is evaluated using real-world logistics drone trajectory data from Shenzhen, China. From two typical scenarios, we can analyze two risk changes during drone operation. Sensitivity analysis reveals how flight risks change as influencing factors change. We also set up model comparisons to examine the characteristics and applicability of our model. Finally, through large-scale data experiments, the high-risk operation areas and periods of logistics drones discovered. These insights offer practical suggestions for ensuring the safe operation of logistics drones, thereby enhancing the efficiency of the system and public acceptance and satisfaction with logistics drones.