基于神经网络模糊控制的单目无人机视觉避障方法研究  

Research on Visual Obstacle Avoidance Method of Monocular UAV Based on Neural Network Fuzzy Control

作　　者：张晓军[1] ZHANG Xiaojun(Huali College Guangdong University of Technology,Guangzhou 511325,China)

机构地区：[1]广东工业大学华立学院,广东广州511325

出　　处：《机械与电子》2020年第6期41-45,共5页Machinery & Electronics

摘　　要：为了提高单目无人机航行安全性能,提出基于神经网络模糊控制的单目无人机视觉避障方法。结合SURF算法采集单目无人机运行路径,优化无人机规避功能模块的功能,规范无人机航行参数标定数值。根据采集到的参数数值,选择并筛选单目无人机航行路径。计算单目无人机障碍物避让路线数值,选择最优避障路线,实现单目无人机视觉避障方案的准确选择,保证运行安全和稳定。通过实验证实,基于神经网络模糊控制的单目无人机视觉避障方法具有较高的避障有效性和实用性,可以更好地保障无人机航行安全,充分满足研究要求。In order to improve the navigation safety performance of monocular UAV,a visual obstacle avoidance method of monocular UAV based on neural network fuzzy control is proposed.Combined with SURF algorithm to collect single-person operation path,optimized UAV avoidance function module function,and standardized UAV navigation parameter calibration value.According to the collected parameters,the flight path of monocular UAV is selected and screened.The numerical value of obstacle avoidance route of monocular UAV is calculated,and the optimal obstacle avoidance route is selected to realize the accurate selection of visual obstacle avoidance scheme of monocular UAV,so as to ensure safe and stable operation.Through experiments,it is proved that the visual obstacle avoidance method of monocular UAV based on neural network fuzzy control has high obstacle avoidance effectiveness and practicability,which can better guarantee the navigation safety of UAV and fully meet the research requirements.

关 键 词：神经网络 模糊控制 单目无人机 视觉避障 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] V279[自动化与计算机技术—控制科学与工程]