| 注册
首页|期刊导航|无线电工程|基于神经网络预测的无人机三维路径规划算法

基于神经网络预测的无人机三维路径规划算法

许志宣 李大鹏 栗保明

无线电工程2026,Vol.56Issue(7):1177-1184,8.
✕
无线电工程2026,Vol.56Issue(7):1177-1184,8.DOI:10.3969/j.issn.1003-3106.2026.07.005

基于神经网络预测的无人机三维路径规划算法

Research on 3D Path Planning Algorithm for Unmanned Aerial Vehicles Based on Neural Network Prediction

许志宣 1李大鹏 1栗保明2

作者信息

  • 1. 南京邮电大学 通信与信息工程学院,江苏 南京 210003
  • 2. 南京理工大学 瞬态物理全国重点实验室,江苏 南京 210003
  • 折叠

摘要

Abstract

Existing sampling-based path planning algorithms have low dependence on environmental information,which tends to result in a large number of sampling points and iterations.However,this environmental information is vast and valuable enough to be analyzed to predict sampling nodes.This algorithm combines the Neural Rapidly-exploring Random Tree Star(NRRT*)algorithm with the Transformer framework to address the issues of traditional algorithms,therefore,a novel path planning algorithm,called TNRRT*algorithm such as a large number of sampling points,iterations,and long path lengths.Specifically,this algorithm uses the NRRT*algorithm as the main framework for sampling and employs a hybrid sampling mode combining uniform sampling and the Transformer model to predict sampling.Extensive experimental results demonstrate that compared to the traditional Rapidly-exploring Random Tree Star(RRT*)algorithm,NRRT* algorithm,and Neural Informed RRT*(NIRRT*)algorithm,the proposed TNRRT* algorithm exhibits superior performance in terms of the number of sampling points,iteration count,and path length.

关键词

路径规划/模型预测/无人机/贝塞尔曲线/快速探索随机树

Key words

path planning/model prediction/unmanned aerial vehicle/Bézier curve/rapidly-exploring random trees

分类

信息技术与安全科学

引用本文复制引用

许志宣,李大鹏,栗保明..基于神经网络预测的无人机三维路径规划算法[J].无线电工程,2026,56(7):1177-1184,8.

基金项目

国家自然科学基金(U24B20187)National Natural Science Foundation of China(U24B20187) (U24B20187)

无线电工程

1003-3106

访问量0
|
下载量0
段落导航相关论文