A multi-unmanned aerial vehicle trajectory optimization and resource allocation method based on diffusion model

Abstract

The present application relates to the technical field of unmanned aerial vehicle trajectory optimization, and particularly relates to a multi-unmanned aerial vehicle trajectory optimization and resource allocation method based on a diffusion model, comprising: constructing a low-altitude unmanned aerial vehicle network based on task execution and communication requirements of multiple unmanned aerial vehicles; decomposing a task scheduling optimization target based on a Lyapunov drift-penalty function; inputting trajectory variables of each unmanned aerial vehicle into a preset diffusion model, and generating an unmanned aerial vehicle optimization trajectory meeting preset flight safety constraints based on a gradient auxiliary guidance mechanism; jointly optimizing running state parameters of each unmanned aerial vehicle in the initialized low-altitude unmanned aerial vehicle network through the unmanned aerial vehicle optimization trajectory; and updating parameters of the preset diffusion model based on a diffusion model loss function after each time slot ends. The present application can deduce a mathematical upper bound of a task queue state and long-term system energy consumption of a multi-unmanned aerial vehicle system and determine a single time slot, thereby improving the precision of multi-unmanned aerial vehicle trajectory generation and resource allocation.

Images