Unmanned aerial vehicle pseudo path planning method based on deep reinforcement learning

Abstract

The invention discloses an unmanned aerial vehicle pseudo path planning method based on deep reinforcement learning, which comprises the steps of: dividing boundary coordinates of a no-fly area on a flight map and marking a start coordinate position and an end coordinate position of a flight task of an unmanned aerial vehicle; sensing the current environment state of the unmanned aerial vehicle before executing the flight task, and selecting a deflection angle and a flight action in the current environment according to an obtained Q function value by using a deep reinforcement learning algorithm; updating a Q function by the unmanned aerial vehicle according to reward return obtained by interaction between flight position data and the environment, wherein the flight position data is continuously received by the unmanned aerial vehicle in the flight process and is provided by ground base station transmitting equipment; judging whether the unmanned aerial vehicle flies according to a preset air route by using the no-fly area as a virtual barrier in the flight process; and if the unmanned aerial vehicle approaches the edge of the no-fly area, guiding the unmanned aerial vehicle to plan a pseudo navigation path through a reward function so as to avoid the no-fly area. The unmanned aerial vehicle pseudo path planning method of the invention realizes the pseudo path planning of the unmanned aerial vehicle in an unknown environment, and improves the intelligence and the safety of the flight of the unmanned aerial vehicle.

Description

technical field [0001] The invention belongs to the technical field of machine learning, and in particular relates to a method for false path planning of unmanned aerial vehicles based on deep reinforcement learning. Background technique [0002] With the considerable progress in the field of computing and artificial intelligence, UAVs are used in more and more fields, especially in the field of military aviation, and the types of tasks performed by UAVs are becoming more and more complex. The fields of military reconnaissance and air transport have played an important role. The requirements for intelligent UAV trajectory planning are also getting higher and higher. When UAVs perform special tasks, UAVs must avoid normal civil aviation flight areas and Radar monitoring area, so as not to interfere with the flight of civil aviation aircraft and radar monitoring. In order to better serve applications in various fields, the research on UAV false path planning has become a res...

AI Technical Summary

Problems solved by technology

[0004] In the process of actually solving the UAV track path planning, according to different tasks and the complexity of the terrain environment, choose an intelligent algorithm that conforms to the track planning. The existing algorithm is based on the unmanned However, in actual situations, some no-fly areas in the airspace are undetectable invisible obstacles, and it is easy to mistakenly fly into the restricted area during the flight of the drone, causing other airspace flight hazards

Images