Abstract: The present disclosure relates to a method for stochastic inspections on power grid lines based on unmanned aerial vehicle-assisted edge computing. According to the method, a stochastic distributed inspection unmanned aerial vehicle is adopted to acquire video images on a target power grid area, which can reduce funds and time costs of inspections. With assistance of superior unmanned aerial vehicle, a goal is to minimize energy consumption of an unmanned aerial vehicle system and extend operation time of the unmanned aerial vehicles under same payload conditions, while processing video image data collected from the inspection unmanned aerial vehicles. The near-far effect generated by communications between mobile unmanned aerial vehicles is eliminated by introducing a NOMA, and position coordinates, system resource allocations and task offload decision schemes are solved by using a method of combining a DDPG algorithm in a Deep reinforcement learning with a genetic algorithm.

Abstract: The present disclosure relates to a method for stochastic inspections on power grid lines based on unmanned aerial vehicle-assisted edge computing. According to the method, a stochastic distributed inspection unmanned aerial vehicle is adopted to acquire video images on a target power grid area, which can reduce funds and time costs of inspections. With assistance of superior unmanned aerial vehicle, a goal is to minimize energy consumption of an unmanned aerial vehicle system and extend operation time of the unmanned aerial vehicles under same payload conditions, while processing video image data collected from the inspection unmanned aerial vehicles. The near-far effect generated by communications between mobile unmanned aerial vehicles is eliminated by introducing a NOMA, and position coordinates, system resource allocations and task offload decision schemes are solved by using a method of combining a DDPG algorithm in a Deep reinforcement learning with a genetic algorithm.