NUAA wins nine first prizes at the national finals of the eighth China Graduate Artificial Intelligence Innovation Competition. [Photo/en.nuaa.edu.cn]
Graduate students from Nanjing University of Aeronautics and Astronautics (NUAA) won nine first prizes at the national finals of the eighth China Graduate Artificial Intelligence Innovation Competition, held from Oct 23 to 25.
The university received more first prizes than any other participating university nationwide, with projects applying AI to challenges ranging from satellite coordination to civil aviation operations.
One award-winning project focuses on autonomous decision-making and coordination for satellite clusters. The team developed a large-scale satellite formation planning algorithm using Lie group-based deep reinforcement learning. Even with limited local information sharing, the system achieved a success rate of up to 97 percent in coordinating large-scale satellite formations.
Another first-prize project developed an intelligent platform to identify the causes of flight delays. Traditionally, determining the cause of a delay has relied heavily on human experience, with responsibility sometimes difficult to establish.
The NUAA team developed an algorithm using multi-level perturbation adversarial training to identify the causes and stages of delays while allowing the underlying evidence to be traced. The platform has been tested using real operational data at Shanghai Pudong International Airport. Its rule-based system matched the airport's findings on the primary cause of delays 84.24 percent of the time, while its AI model achieved accuracy rates of 95.50 percent in identifying the causes of delays and 95.86 percent in identifying the relevant stage of operations.
