Complex Network-Based Airspace Risk Analysis Under Autonomous Operation Mode of Aircraft

In order to meet the refined requirements of airspace risk analysis in autonomous operation mode, this paper proposes a flight conflict network model that integrates the spatial position proximity and velocity obstacle relationship between aircraft. Firstly, regarding the spatial position relationship, considering the uncertainty during the trajectory operation, the CNS performance factors affecting the trajectory error are modeled into the traditional conflict probability calculation model, and the spatial position proximity is determined by calculating the conflict probability between aircrafts; secondly, regarding the velocity obstacle relationship, considering the relative position and relative velocity relationship between aircraft, the conflict tendency between aircraft in future moments is determined based on the three-dimensional velocity obstacle model; then, aircraft that satisfy both position proximity and velocity obstacle relationships are connected, and network edge weights are determined; finally, risk assessment indicators are established based on complex network theory, the impact of different network construction methods on airspace complexity is compared and analyzed, and airspace operation risk levels are classified by combining with risk vector. The results show that: the conflict probability model based on CNS performance can realize the fine discrimination of aircraft spatial position proximity by combining the operational state parameters of the aircraft; the proposed network construction method can finely evaluate the risk between aircrafts, which reduces 85.87 %, 66.67 %, and 64.86 % of the network edges compared to the other three flight conflict networks, and demonstrates significant advantages in terms of network nodes and overall performance indicators; the risk vector based on the network topology indicators is effective in classifying the airspace operation risk level.