Intelligent obstacle avoidance method for unmanned aerial vehicle based on autonomous learning

Abstract

The invention discloses an intelligent obstacle avoidance method for an unmanned aerial vehicle based on autonomous learning, and so, the unmanned aerial vehicle intelligently and autonomously learnsby utilizing visual information obtained by a camera in a three-dimensional visualization simulation environment and according to terrain height and obstacle height, constant altitude flight control is performed on the unmanned aerial vehicle by utilizing a trained network model as an intelligent body, the flight height is adjusted in real time, and an automatic terrain following application is achieved to complete an autonomous obstacle avoidance task. The method creates the three-dimensional visualization simulation environment, provides a good training environment for an intelligent autonomous obstacle avoidance algorithm, implements an interaction interface for human-machine real-time operation, and provides the possibility for a transfer training from a simulation environment to a true environment. The method provides a simulation test platform for other deep reinforcement learning algorithms, and is convenient for the intelligent body to perform deep reinforcement training and testing of multiple scenarios, different tasks, and multiple algorithms.

Description

【Technical field】 [0001] The invention belongs to the technical field of intelligent control and guidance systems and methods for autonomous obstacle avoidance technology of unmanned aerial vehicles, and in particular relates to an intelligent obstacle avoidance method of unmanned aerial vehicles based on autonomous learning. 【Background technique】 [0002] Obstacle avoidance ability is the key link to realize the automation and even intelligence of UAVs. A perfect UAV autonomous obstacle avoidance system can avoid obstacles in the flight path in time, greatly reducing the damage and damage of UAVs caused by operation errors. The incidence of accidents involving people and structures. [0003] Autonomous obstacle avoidance flight is very important for UAVs, which can ensure that UAVs can complete complex, multi-functional and difficult maneuvers. The development of UAV obstacle avoidance technology can be divided into three stages, one is to sense obstacles and hover; the o...

AI Technical Summary

Problems solved by technology

However, this algorithm needs to manually specify the eigenvalues ​​to be extracted in terms of image processing, and is easily affected by factors such as illumination and obstacle positions, and has poor robustness.

The general deep learning method allows the UAV to learn the characteristics of obstacles by training a large number of perceptual images. However, it is difficult to clarify the boundaries of obstacles and lack the decision-making ability to avoid obstacles correctly.

Images