A novel multi-objective trajectory optimization approach for drone swarms

This paper aims at the problems of difficult mobility prediction, rigid topology and resource conflict faced by dynamic path planning in unmanned aerial vehicle-assisted mobile edge computing, and proposes a trinity framework integrating environmental perception, topology optimization and resource collaboration. The accuracy of equipment trajectory prediction is improved through the terrain coupling displacement model, the time-varying weight function is designed to achieve dynamic network reconstruction, and the three-dimensional optimization model of energy consumption-delay-safety is established. Experiments show that this framework reduces the average task delay to 4.2ms (96.5% of tasks <13ms), the delay increase under attack is only 10.5%, and the dispersion of service node distribution is reduced by 42%, providing a theoretically complete solution for the 6G air-space-ground cooperative network.