Özet:

Abstract Recent developments in Unmanned Aerial Vehicle (UAV) technology have shown that they will form an integral element of future communication and networking infrastructure. Although several studies have offered UAV-assisted methods for enhancing the performance of existing networks by increasing coverage and capacity but the architectures of autonomous UAV networks based on artificial intelligence has not yet been thoroughly investigated. However, the most current models for logistics UAV delivery do not account for the energy consumption of logistics UAVs or the varying schedules of their clients, meaning they are not applicable to real-world transportation networks. As a result, for a smart transportation system, we suggest reducing the overall energy cost of various logistics UAVs throughout the time it takes to deliver individual items. In this research, we maximize the UAV power by posing the UAV path planning issue as a traveling salesman problem. The UAV route planning is optimized under the restrictions of node energy consumption and task deadlines to achieve maximum energy efficiency of cooperative computing over the course of a UAV's life cycle. A Deep Q Network (DQN) based path planning algorithm is suggested to adjust to the uncertain and changing environment over time. In comparison to other algorithms, the proposed one performs better in simulations, increases the computational productivity of dynamic computing by a large margin, and achieves a good equilibrium between the two energy inputs. We also think about minimizing the UAV's spin rate to maximize efficiency and decrease power consumption. By lowering the number of turns while still visiting all of the waypoints, our suggested technique uses 2-5 times less energy.

Anahtar Kelimeler: